--- title: "rocbc" author: "LE Bantis, B Brewer, CT Nakas, B Reiser" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{rocbc} %\VignetteEngine{knitr::knitr} %\VignetteEncoding{UTF-8} --- ```{r eval=TRUE, error=FALSE, warning=FALSE, cache=FALSE, comment=FALSE, echo=FALSE, message=FALSE, results=FALSE} # Hello, world! # # This is an example function named 'hello' # which prints 'Hello, world!'. # # You can learn more about package authoring with RStudio at: # # http://r-pkgs.had.co.nz/ # # Some useful keyboard shortcuts for package authoring: # # Install Package: 'Ctrl + Shift + B' # Check Package: 'Ctrl + Shift + E' # Test Package: 'Ctrl + Shift + T' library(pracma) library(clinfun) library(splancs) library(mvtnorm) library(matlab) library(formattable) library(pROC) library(MRMCaov) ``` ```{r, echo=FALSE, eval=TRUE} rocboxcox<-function(marker, D, alpha, plots, printProgress = FALSE){ if (plots!="on"){plots="off"} if (length(marker) != length(D)) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (sum(is.na(marker)) > 0 | sum(is.na(D)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if (alpha <= 0 | alpha >= 1) { stop("ERROR: The level of significance, alpha, should be set between 0 and 1. A common choice is 0.05.") } else if (min(marker<= 0)) { stop("ERROR: All marker scores need to be positive") } else if (sum(marker < 0) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { x=marker[D==0] y=marker[D==1] if (plots!="on"){plots="off"} n1=length(x) n2=length(y) xor=x;yor=y; Za=qnorm(1-alpha/2) likbox<-function(x,y,h){ n=length(x); m=length(y); out=c(); for (i in 1:length(h)){ #print(i) if (h[i]==0){ xh=log(x); yh=log(y); } else { xh=((x^h[i])-1)/h[i]; yh=((y^h[i])-1)/h[i]; } out[i]<-c( -n/2*log(sum((xh-sum(xh)/n)^2)/n) -m/2*log(sum((yh-sum(yh)/m)^2)/m) +(h[i]-1)*(sum(log(x))+sum(log(y)))) } return(out) } boxcoxleo<-function(x,y){ init=1; logL<-function(h){ -likbox(x,y,h) } #lam=fminsearch(logL,init) #lam=c(lam$optbase$xopt) lam<- optim(1,logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) transx=((x^lam)-1)/lam transy=((y^lam)-1)/lam #test1=print(shapiro.test(transx)) #test2=print(shapiro.test(transy)) return(list(transformation.parameter=lam,transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam )) } boxcoxleo2<-function(x,y){ init=1; logL<-function(h){ -likbox(x,y,h) } lam<- optim(1,logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) #lam=fminsearch(logL,init) #lam=c(lam$optbase$xopt) transx=((x^lam)-1)/lam transy=((y^lam)-1)/lam return(list(transformation.parameter=lam,transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam )) } COVlam <-function (x,y){ #%Accepts Xg(1) # %sh-->g(2) # %md-->g(3) # %sd-->g(4) # %lam-->g(5) # % logL=@(g) -n*log(sh)-1/(2*sh^2)*sum((xlam-mh)^2)+(lam-1)*sum(log(x)) - # % m*log(sd)-1/(2*sd^2)*sum((ylam-md)^2)+(lam-1)*sum(log(y)) outt=boxcoxleo(x,y); #the original scores transx=outt$transx transy=outt$transy lam=outt$transformation.parameter mh=mean(transx); xlam=transx; md=mean(transy); ylam=transy; sh=sqrt(1/(length(transx))*sum((transx-mean(transx))^2)); sd=sqrt(1/(length(transy))*sum((transy-mean(transy))^2)); I=zeros(5,5); I[1,1]=n/sh^2; I[2,2]=-(n/sh^2-3/sh^4*sum((xlam-mh)^2)); I[3,3]=m/sd^2; I[4,4]=-(m/sd^2-3/sd^4*sum((ylam-md)^2)); kk= sum(((mh - (x^lam - 1)/lam)*((2*(x^lam - 1))/lam^3 - (2*x^lam*log(x))/lam^2 + (x^lam*log(x)^2)/lam))/sh^2) + - sum(((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)^2/sd^2) + - sum(((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)^2/sh^2) + + sum(((md - (y^lam - 1)/lam)*((2*(y^lam - 1))/lam^3 - (2*y^lam*log(y))/lam^2 + (y^lam*log(y)^2)/lam))/sd^2); I[5,5]=-kk ; I[1,5]=sum(((2*(x^lam - 1))/lam^2 - (2*x^lam*log(x))/lam)/(2*sh^2)); I[2,5]=-sum((2*((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)*(mh - (x^lam - 1)/lam))/sh^3); I[3,5]=-sum(-((2*(y^lam - 1))/lam^2 - (2*y^lam*log(y))/lam)/(2*sd^2)); I[4,5]=-sum((2*((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)*(md - (y^lam - 1)/lam))/sd^3); I[5,1]=I[1,5]; I[5,2]=I[2,5]; I[5,3]=I[3,5]; I[5,4]=I[4,5]; S= inv(I); #%that's better return(list(out=S)) } cc=boxcoxleo(x,y) transx=cc$transx transy=cc$transy #x11() #qqnorm(x) #title(main=" for X") #x11() #qqnorm(y) #title(main=" for Y") roc<-function(t,trans_x,trans_y){ na = length(trans_x) nb = length(trans_y) stransx=sqrt(var(trans_x)*(na-1)/na) stransy=sqrt(var(trans_y)*(nb-1)/nb) 1-pnorm(qnorm(1-t, mean=mean(trans_x), sd=stransx), mean=mean(trans_y), sd=stransy) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } if (plots=="on"){ txt <- paste("Box-Cox Based ROC, alpha =", formattable(alpha, digits = 3)) plot(linspace(0,1,1000),roc(linspace(0,1,1000),transx,transy),main=" ",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="red") lines(linspace(0,1,10),linspace(0,1,10),type="l", lty=2) title(main=txt) } rocfun=function(t){1-pnorm(qnorm(1-t,mean=mean(transx),sd=sqrt(var(transx)*(length(transx)-1)/length(transx))),mean=mean(transy),sd=sqrt(var(transy)*(length(transy)-1)/length(transy)))} m1hat=mean(transx) m2hat=mean(transy) s1hat=sqrt(var(transx)*(length(transx)-1)/length(transx)) s2hat=sqrt(var(transy)*(length(transy)-1)/length(transy)) n1=length(x) n2=length(y) #NOTE THAT: m1hat->m1 # m2hat->m2 # s1hat->s1hat # s2hat->s2hat ######################################### derde =c( (s2hat^2 + (s1hat*s2hat*(2*m1hat - 2*m2hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s2hat*(s1hat^2 - s2hat^2)), ((s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2) - 2*m2hat*s1hat + (s1hat*s2hat*(2*s1hat*log(s1hat^2/s2hat^2) + (2*(s1hat^2 - s2hat^2))/s1hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2) - (2*s1hat*(m1hat*s2hat^2 - m2hat*s1hat^2 + s1hat*s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2)^2)/s2hat, -((s1hat^2 + (s1hat*s2hat*(2*m1hat - 2*m2hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2) - 1)/s2hat, ((2*m1hat*s2hat + s1hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2) - (s1hat*s2hat*(2*s2hat*log(s1hat^2/s2hat^2) + (2*(s1hat^2 - s2hat^2))/s2hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2) + (2*s2hat*(m1hat*s2hat^2 - m2hat*s1hat^2 + s1hat*s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2)^2)/s2hat - (m2hat + (m1hat*s2hat^2 - m2hat*s1hat^2 + s1hat*s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2))/(s1hat^2 - s2hat^2))/s2hat^2); derdp =c( -((s2hat^2 + (s1hat*s2hat*(2*m1hat - 2*m2hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2) + 1)/s1hat, (m1hat + (m1hat*s2hat^2 - m2hat*s1hat^2 + s1hat*s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2))/(s1hat^2 - s2hat^2))/s1hat^2 - ((s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2) - 2*m2hat*s1hat + (s1hat*s2hat*(2*s1hat*log(s1hat^2/s2hat^2) + (2*(s1hat^2 - s2hat^2))/s1hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2) - (2*s1hat*(m1hat*s2hat^2 - m2hat*s1hat^2 + s1hat*s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2)^2)/s1hat, (s1hat^2 + (s1hat*s2hat*(2*m1hat - 2*m2hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat*(s1hat^2 - s2hat^2)), -((2*m1hat*s2hat + s1hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2) - (s1hat*s2hat*(2*s2hat*log(s1hat^2/s2hat^2) + (2*(s1hat^2 - s2hat^2))/s2hat))/(2*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2) + (2*s2hat*(m1hat*s2hat^2 - m2hat*s1hat^2 + s1hat*s2hat*(log(s1hat^2/s2hat^2)*(s1hat^2 - s2hat^2) + (m1hat - m2hat)^2)^(1/2)))/(s1hat^2 - s2hat^2)^2)/s1hat); outt=COVlam(xor,yor); S=outt$out; S=S[1:4,1:4]; # if (sum(is.infinite(S)) > 0 | # sum(is.nan(S)) > 0 | # sum(is.na(S)) > 0 | # sum(diag(S) < 0) > 0) { # stop("ERROR: The information matrix cannot be inverted - this might be due to the scale of the marker. Try rescaling the marker measurements by subtracting a constant or by dividing with a constant all marker measurements (of both groups) before you try again.") # } varde=t(derde)%*%S%*%t(t(derde)); vardp=t(derdp)%*%S%*%t(t(derdp)); covdedp=t(derde)%*%S%*%t(t(derdp)); ########################################## vardeltase=varde; vardeltasp=vardp; ######################### cutoff= (s1hat^2*m2hat-s2hat^2*m1hat-s1hat*s2hat*sqrt((m1hat-m2hat)^2+(s1hat^2-s2hat^2)*log(s1hat^2/s2hat^2)))/(s1hat^2-s2hat^2) deltasp= (((s1hat^2*m2hat-s2hat^2*m1hat-s1hat*s2hat*sqrt((m1hat-m2hat)^2+(s1hat^2-s2hat^2)*log(s1hat^2./s2hat^2)))/(s1hat^2-s2hat^2))-m1hat)/(s1hat) deltase= (m2hat-((s1hat^2*m2hat-s2hat^2*m1hat-s1hat*s2hat*sqrt((m1hat-m2hat)^2+(s1hat^2-s2hat^2)*log(s1hat^2/s2hat^2)))/(s1hat^2-s2hat^2)))/s2hat Se=pnorm(deltase) Sp=pnorm(deltasp) Se Sp if (plots=="on"){ points(1-Sp,Se, col = "red") } covdeltas = matrix( c(vardeltasp, covdedp, covdedp, vardeltase), nrow=2, ncol=2, byrow = TRUE) # fill m # if (sum(is.infinite(covdeltas)) > 0 | # sum(is.nan(covdeltas)) > 0) { # stop("ERROR: The information matrix cannot be inverted - this might be due to the scale of the marker. Try rescaling the marker measurements by subtracting a constant or by dividing with a constant all marker measurements (of both groups) before you try again.") # } svdA=svd(inv(covdeltas)) a=sqrt(qchisq(1-alpha,2))/sqrt(svdA$d[1]) b=sqrt(qchisq(1-alpha,2))/sqrt(svdA$d[2]) theta=seq(from=0,to=2*pi+1/40,by=1/40) state1=a*cos(theta) state2=b*sin(theta) states=rbind(state1,state2) X=svdA$v %*%states x1=X[1,]+deltasp x2=X[2,]+deltase px1=pnorm(x1); px2=pnorm(x2); yegg=px2 xegg=1-px1 if (plots=="on"){ lines(xegg,yegg,col="green") } #----Area egg--------------- bxegg=c(xegg,xegg[1]) byegg=c(yegg,yegg[1]) ell <- cbind(bxegg, byegg) areaegg=areapl(ell) #----End Area egg--------------- margcisp=c(pnorm(deltasp-qnorm(1-alpha/2)*sqrt(vardeltasp)), pnorm(deltasp+qnorm(1-alpha/2)*sqrt(vardeltasp))) margcise=c(pnorm(deltase-qnorm(1-alpha/2)*sqrt(vardeltase)), pnorm(deltase+qnorm(1-alpha/2)*sqrt(vardeltase))) #-----Now deal with the rectangle----- cisp=c(deltasp-qnorm(1-alpha/4)*sqrt(vardeltasp), deltasp+qnorm(1-alpha/4)*sqrt(vardeltasp)) cise=c(deltase-qnorm(1-alpha/4)*sqrt(vardeltase), deltase+qnorm(1-alpha/4)*sqrt(vardeltase)) bsenslb=pnorm(cise[1])#exp(cibootsse[1])/(1+exp(cibootsse[1])) #NOW trans back, going to construct a 95% rectangle for boots bsensub=pnorm(cise[2])#exp(cibootsse[2])/(1+exp(cibootsse[2])) #NOW trans back, going to construct a 95% rectangle for boots bspeclb=pnorm(cisp[1])#exp(cibootssp[1])/(1+exp(cibootssp[1])) #NOW trans back, going to construct a 95% rectangle for boots bspecub=pnorm(cisp[2])#exp(cibootssp[2])/(1+exp(cibootssp[2])) #NOW trans back, going to construct a 95% rectangle for boots #-----End of the rectangle------------ if (plots=="on"){ #-----Start--plot---Rectangle-------------------- lines(c(1-bspeclb,1-bspecub),c(bsenslb,bsenslb),col="black") lines(c(1-bspecub,1-bspecub),c(bsenslb,bsensub),col="black") lines(c(1-bspecub,1-bspeclb),c(bsensub,bsensub),col="black") lines(c(1-bspeclb,1-bspeclb),c(bsenslb,bsensub),col="black") #-----End--plot---Rectangle-------------------- } brectx=c(1-bspeclb,1-bspecub, 1-bspecub,1-bspecub,1-bspecub,1-bspeclb,1-bspeclb,1-bspeclb) brecty=c(bsenslb,bsenslb ,bsenslb,bsensub, bsensub,bsensub ,bsenslb,bsensub) brectx=c(brectx, brectx[1]); brecty=c(brecty, brecty[1]); re <- cbind(brectx, brecty) arearect=areapl(re) #line for the Youden: if (plots=="on"){ lines(c(1-Sp,1-Sp),c(rocuseless(1-Sp),Se),col="blue") } #================AUC================ invAUC= ((m2hat-m1hat)/s2hat)/(sqrt(1+(s1hat/s2hat)^2)); auc=pnorm(invAUC) #====================AUC1============================== dm1 =-1/(s2hat*sqrt(1+(s1hat/s2hat)^2)); ds1 = (s1hat*(m1hat-m2hat))/(s2hat^3*(1+(s1hat/s2hat)^2)^(3/2)) dm2 =1/(s2hat*sqrt(1+(s1hat/s2hat)^2)); ds2 =((m1hat-m2hat))/(s2hat^2*(1+(s1hat/s2hat)^2)^(3/2)) I=zeros(5,5); sh=sqrt(1/(length(transx))*sum((transx-mean(transx))^2)) sd=sqrt(1/(length(transy))*sum((transy-mean(transy))^2)) mh=m1hat; md=m2hat; n=n1; m=n2; xlam=transx; ylam=transy; lam=cc$transformation.parameter I[1,1]=n/sh^2; I[2,2]=-(n/sh^2-3/sh^4*sum((xlam-mh)^2)); I[3,3]=m/sd^2; I[4,4]=-(m/sd^2-3/sd^4*sum((ylam-md)^2)); kk= sum(((mh - (x^lam - 1)/lam)*((2*(x^lam - 1))/lam^3 - (2*x^lam*log(x))/lam^2 + (x^lam*log(x)^2)/lam))/sh^2) + - sum(((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)^2/sd^2) + - sum(((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)^2/sh^2) + + sum(((md - (y^lam - 1)/lam)*((2*(y^lam - 1))/lam^3 - (2*y^lam*log(y))/lam^2 + (y^lam*log(y)^2)/lam))/sd^2); I[5,5]=-kk ; I[1,5]=sum(((2*(x^lam - 1))/lam^2 - (2*x^lam*log(x))/lam)/(2*sh^2)); I[2,5]=-sum((2*((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)*(mh - (x^lam - 1)/lam))/sh^3); I[3,5]=-sum(-((2*(y^lam - 1))/lam^2 - (2*y^lam*log(y))/lam)/(2*sd^2)); I[4,5]=-sum((2*((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)*(md - (y^lam - 1)/lam))/sd^3); I[5,1]=I[1,5] I[5,2]=I[2,5] I[5,3]=I[3,5] I[5,4]=I[4,5] S=inv(I) S=S[1:4,1:4] ## EVERYTHING ELSE ================================================================================ varauc=t(c(dm1,ds1,dm2,ds2))%*% S %*% t(t(c(dm1,ds1,dm2,ds2))) CIauc=c(invAUC-Za*sqrt(varauc), invAUC+Za*sqrt(varauc)) CIauc=pnorm(CIauc) CIauc Zauc=invAUC/sqrt(varauc) pvalauc=2*pnorm(-abs(Zauc)) #================= DeltaBClamJT================ mx=m1hat; my=m2hat; sx=s1hat; sy=s2hat; c=cutoff; b=sy/sx; a=my-mx; Jhat=pnorm((m2hat-cutoff)/s2hat)+pnorm((cutoff-m1hat)/s1hat)-1; radd=a^2+(b^2-1)*sx^2*log(b^2); zy=(my-c)/sy; zx=(c-mx)/sx; dcdmx= (b^2 + a*b*radd^(-1/2)*(-1))/(b^2-1); dcdsx= (-2*a*b^2)/((b^2-1)^2*sx) + (((b*(b^2+1))*(radd)^(1/2))/((b^2-1)^2*sx) - (sx*b*radd^(-1/2))/(b^2-1)*(log(b^2)+b^(2)-1) ); dcdmy= (-1 + a*b*radd^(-1/2))/(b^2-1); dcdsy= (2*a*b)/((b^2-1)^2*sx) + (((-b^2-1)*(radd)^(1/2))/((b^2-1)^2*sx) + (sy*b*radd^(-1/2))/(b^2-1)*(log(b^2)+1-b^(-2)) ); d1=-sx^(-1)*dnorm(zx)+dcdmx*(sx^(-1)*dnorm(zx)-(sy^(-1)*dnorm(zy) )); d2=-zx*sx^(-1)*dnorm(zx)+dcdsx*(sx^(-1)*dnorm(zx)-(sy^(-1)*dnorm(zy) )) ; d3=sy^(-1)*dnorm(zy)+dcdmy*(sx^(-1)*dnorm(zx)-(sy^(-1)*dnorm(zy) )) ; d4=-zy*sy^(-1)*dnorm(zy)+dcdsy*(sx^(-1)*dnorm(zx)-(sy^(-1)*dnorm(zy) )); #varJ=t(c(d1, d2, d3, d4))%*%S%*%[d1 d2 d3 d4]'; varJ=t(c(d1,d2,d3,d4))%*% S %*% t(t(c(d1,d2,d3,d4))) # Jstar=(Jhat+1)/2; # Jstar=log(Jstar/(1-Jstar)); # VarJstar=(4/(Jhat^2 - 1)^2)*varJ Jstar=qnorm(Jhat); varJstar=(1/(dnorm(qnorm(Jhat))))^2*varJ ZJstar=Jstar/sqrt(varJstar) pvalJ=2*pnorm(-abs(ZJstar)) CIstar=c(Jstar-Za*sqrt(varJstar), Jstar+Za*sqrt(varJstar)); CI=c(0,0) #CI[1]=exp(CIstar[1])/(1+exp(CIstar[1]))*2-1 #CI[2]=exp(CIstar[2])/(1+exp(CIstar[2]))*2-1 CI[1]=pnorm(CIstar[1]) CI[2]=pnorm(CIstar[2]) CIJ=CI; Jhat CIwidth=max(CI)-min(CI) boots=0;c1hat_boots=0;c1hat_boots_or=0; c1hat=cutoff; #===========CIs for the cutoff================== for (boots in 1:1000){ #if (boots>=2){print((m1hat_boots*(b^2-1)-a+b*sqrt(a^2+(b^2-1)*s1hat_boots^2*log(b^2)))/(b^2-1))} #tryCatch({ atx = sample(1:n1,n1,replace=T) aty = sample(1:n2,n2,replace=T) if (printProgress) { if (boots==100){print(boots);print("bootstap samples out of 1000...")} if (boots==200){print(boots);print("bootstap samples out of 1000...")} if (boots==300){print(boots);print("bootstap samples out of 1000...")} if (boots==400){print(boots);print("bootstap samples out of 1000...")} if (boots==500){print(boots);print("bootstap samples out of 1000...")} if (boots==600){print(boots);print("bootstap samples out of 1000...")} if (boots==700){print(boots);print("bootstap samples out of 1000...")} if (boots==800){print(boots);print("bootstap samples out of 1000...")} if (boots==900){print(boots);print("bootstap samples out of 1000...")} if (boots==1000){print(boots);print("Bootstrapping complete. You can request all results through the summary() function.")} } xboots=xor[atx] yboots=yor[aty] ccb=boxcoxleo2(xboots,yboots) lam_boots=ccb$transformation.parameter transx1_boots=ccb$transx; transx2_boots=ccb$transy; m1hat_boots=mean(transx1_boots) m2hat_boots=mean(transx2_boots) s1hat_boots=sqrt(var(transx1_boots)*(length(transx1_boots)-1)/length(transx1_boots)) s2hat_boots=sqrt(var(transx2_boots)*(length(transx2_boots)-1)/length(transx2_boots)) b=s2hat_boots/s1hat_boots a=m2hat_boots-m1hat_boots #print(b) #print(a) c1hat_boots[boots]=(m1hat_boots*(b^2-1)-a+b*sqrt(a^2+(b^2-1)*s1hat_boots^2*log(b^2)))/(b^2-1) kr=(m1hat_boots*(b^2-1)-a+b*sqrt(a^2+(b^2-1)*s1hat_boots^2*log(b^2)))/(b^2-1) c1hat_boots_or[boots] = ((kr*lam_boots+1))^(1/lam_boots) #}, error=function(e){}) } varboots=var(na.omit(c1hat_boots_or)) ccc=((c1hat*lam+1))^(1/lam) CI_BCAN=c(ccc-Za*sqrt(var(na.omit(c1hat_boots_or))), ccc+Za*sqrt(var(na.omit(c1hat_boots_or)))) CIwidth=max(CI_BCAN)-min(CI_BCAN) CIcutoff=CI_BCAN; cutoff=ccc; allc1=c1hat_boots_or #CI_BCPB=c(quantile(na.omit(c1hat_boots_or),0.025,"type"=2), quantile(na.omit(c1hat_boots_or),0.975,"type"=2)) #CIwidth=max(CI_BCPB)-min(CI_BCPB) #CI_BCPB #legend("bottomright", legend=c("ROC estimate", "max of Youden index (J)", "rectangular 95% confidence region of the optimal operating point", "rectangular 95% confidence region of the optimal operating point"), # col=c("red", "blue", "black", "green"), lty=c(1,1,1,1), cex=0.8) if (plots=="on"){ legend("bottomright", legend=c(paste("ROC estimate with AUC = ", formattable(auc, digits = 4, format = "f"), ", CI: (", formattable(CIauc[1], digits = 4, format = "f"), ", ", formattable(CIauc[2], digits = 4, format = "f") ,")", sep = ""), paste("Maximized Youden index = ", formattable(Jhat, digits = 4, format = "f"), ", CI: (", formattable((CIJ[1]), digits = 4, format = "f"), ", ", formattable((CIJ[2]), digits = 4, format = "f"), ")", sep = ""), paste("Optimal pair of (FPR,TPR): (", formattable(1-Sp, digits = 4, format = "f"), ", ", formattable(Se, digits = 4, format = "f"), ")", sep = ""), paste("Area of the rect. conf. region =", formattable(arearect, digits = 4, format = "f")), paste("Area of the egg conf. region =", formattable(areaegg, digits = 4, format = "f")), paste("Youden based cutoff: ", formattable(cutoff, digits = 4, format = "f"),", CI: (", formattable(CIcutoff[1], digits = 4, format = "f"), ", ", formattable(CIcutoff[2], digits = 4, format = "f"), ")", sep = ""), paste("Sp Marginal CI: (", c(formattable(margcisp[1], digits = 4, format = "f")), ", ", formattable(margcisp[2], digits = 4, format = "f"), ")", sep = ""), paste("Se Marginal CI: (", c(formattable(margcise[1], digits = 4, format = "f")), ", ", formattable(margcise[2], digits = 4, format = "f"), ")", sep = "")), col=c("red", "blue", "red", "black", "green", "white", "white", "white"), lty=c(1,1,1,1,1,1,NA, NA), pch = c(NA, NA, 19, NA, NA, NA, NA, NA), cex=0.6) } #return(list(transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam , transformation.parameter=lam, AUC=auc, AUCCI=CIauc, J=Jhat, JCI=CIJ, Sens=Se, CImarginalSens=margcise, Spec=Sp, CImarginalSpec=margcisp, cutoff=cutoff, CIcutoff=CIcutoff, areaegg=areaegg, arearect=arearect, mxlam=mean(transx), sxlam=std(transx), mylam=mean(transy), sylam=std(transy))) res <- matrix(c(auc,CIauc[1],CIauc[2], Jhat,CIJ[1],CIJ[2], cutoff,CIcutoff[1],CIcutoff[2], Sp,margcisp[1],margcisp[2], Se,margcise[1],margcise[2], lam,NA,NA),ncol=3,byrow=TRUE) colnames(res) <- c("Estimate","Confidence","Interval") rownames(res) <- c("AUC","Jhat","cutoff","Sp","Se", "Lambda") res <- data.frame(res) res <- formattable(as.matrix(res), digits = 4, format = "f") res if (auc < 0.5) { print("NOTE: AUC < 0.5; the ordering of the two groups may need to be reversed.") } return(list(transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam , transformation.parameter=lam, AUC=auc, AUCCI=CIauc, pvalueAUC=pvalauc, J=Jhat, JCI=CIJ, pvalueJ=pvalJ, Sens=Se, CImarginalSens=margcise, Spec=Sp, CImarginalSpec=margcisp, cutoff=cutoff, CIcutoff=CIcutoff, areaegg=areaegg, arearect=arearect, mxlam=mean(transx), sxlam=sqrt(var(transx)*(length(transx)-1)/length(transx)), mylam=mean(transy), sylam=sqrt(var(transy)*(length(transy)-1)/length(transy)), results=res , rocfun=rocfun)) } } ``` ```{r, echo=FALSE, eval=TRUE} rocboxcoxCI<-function(marker, D, givenSP, givenSE, alpha, plots){ if (plots!="on"){plots="off"} if (length(marker) != length(D)) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (sum(is.na(marker)) > 0 | sum(is.na(D)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if (alpha <= 0 | alpha >= 1) { stop("ERROR: The level of significance, alpha, should be set between 0 and 1. A common choice is 0.05.") } else if (!((sum(is.na(givenSP)) == 0 & sum(is.na(givenSE) > 0)) | (sum(is.na(givenSE)) == 0 & sum(is.na(givenSP) > 0)))) { stop("ERROR: Exactly one of 'givenSP' and 'givenSE' must be set to NA.") } else if (sum(marker < 0) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { #graphics.off() if (plots!="on"){plots="off"} xor=marker[D==0] yor=marker[D==1] x = xor y = yor likbox<-function(x,y,h){ n=length(x); m=length(y); out=c(); for (i in 1:length(h)){ #print(i) if (h[i]==0){ xh=log(x); yh=log(y); } else { xh=((x^h[i])-1)/h[i]; yh=((y^h[i])-1)/h[i]; } out[i]<-c( -n/2*log(sum((xh-sum(xh)/n)^2)/n) -m/2*log(sum((yh-sum(yh)/m)^2)/m) +(h[i]-1)*(sum(log(x))+sum(log(y)))) } return(out) } boxcoxleo<-function(x,y){ init=1; logL<-function(h){ -likbox(x,y,h) } #lam=fminsearch(logL,init) #lam=c(lam$optbase$xopt) lam<- optim(1,logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) transx=((x^lam)-1)/lam transy=((y^lam)-1)/lam #test1=print(shapiro.test(transx)) #test2=print(shapiro.test(transy)) return(list(transformation.parameter=lam,transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam )) } boxcoxleo2<-function(x,y){ init=1; logL<-function(h){ -likbox(x,y,h) } lam<- optim(1,logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) #lam=fminsearch(logL,init) #lam=c(lam$optbase$xopt) transx=((x^lam)-1)/lam transy=((y^lam)-1)/lam return(list(transformation.parameter=lam,transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam )) } cc=boxcoxleo2(x,y) transx=cc$transx transy=cc$transy #x11() #qqnorm(x) #title(main=" for X") #x11() #qqnorm(y) #title(main=" for Y") roc<-function(t,trans_x,trans_y){ na = length(trans_x) nb = length(trans_y) stransx=sqrt(var(trans_x)*(na-1)/na) stransy=sqrt(var(trans_y)*(nb-1)/nb) 1-pnorm(qnorm(1-t, mean=mean(trans_x), sd=stransx), mean=mean(trans_y), sd=stransy) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } # x11() #plot.new() if (plots == "on") { txt <- paste("Box-Cox Based ROC, alpha =", round(alpha,3) ) plot(linspace(0,1,1000),roc(linspace(0,1,1000),transx,transy),main=" ",xlab="FPR",ylab="TPR",type="l",col="red") lines(linspace(0,1,10),linspace(0,1,10),type="l", lty=2) title(main=txt) } m1hat=mean(transx) m2hat=mean(transy) # s1hat=std(transx) # s2hat=std(transy) n1=length(x) n2=length(y) I=zeros(5,5); sh=sqrt(1/(length(transx))*sum((transx-mean(transx))^2)) sd=sqrt(1/(length(transy))*sum((transy-mean(transy))^2)) s1hat = sh s2hat = sd mh=m1hat; md=m2hat; n=n1; m=n2; xlam=transx; ylam=transy; lam=cc$transformation.parameter I[1,1]=n/sh^2; I[2,2]=-(n/sh^2-3/sh^4*sum((xlam-mh)^2)); I[3,3]=m/sd^2; I[4,4]=-(m/sd^2-3/sd^4*sum((ylam-md)^2)); kk= sum(((mh - (x^lam - 1)/lam)*((2*(x^lam - 1))/lam^3 - (2*x^lam*log(x))/lam^2 + (x^lam*log(x)^2)/lam))/sh^2) + - sum(((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)^2/sd^2) + - sum(((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)^2/sh^2) + + sum(((md - (y^lam - 1)/lam)*((2*(y^lam - 1))/lam^3 - (2*y^lam*log(y))/lam^2 + (y^lam*log(y)^2)/lam))/sd^2); I[5,5]=-kk ; I[1,5]=sum(((2*(x^lam - 1))/lam^2 - (2*x^lam*log(x))/lam)/(2*sh^2)); I[2,5]=-sum((2*((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)*(mh - (x^lam - 1)/lam))/sh^3); I[3,5]=-sum(-((2*(y^lam - 1))/lam^2 - (2*y^lam*log(y))/lam)/(2*sd^2)); I[4,5]=-sum((2*((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)*(md - (y^lam - 1)/lam))/sd^3); I[5,1]=I[1,5] I[5,2]=I[2,5] I[5,3]=I[3,5] I[5,4]=I[4,5] S=inv(I) S=S[1:4,1:4] #============================================= #GRID LINES: kk=1;CIROC=c(1,1);CIROCvec1=0;CIROCvec2=0;Sehat=0; tt=c(linspace(0,1,2000)) for (t in tt){ norminvROC= ((m2hat-m1hat)/s2hat+s1hat/s2hat*qnorm(t)) #====================ROC1============================== dm1 =-1/s2hat; ds1 =-(2^(1/2)*erfcinv(2*t))/s2hat; dm2 =1/s2hat; ds2 =(m1hat - m2hat)/s2hat^2 + (2^(1/2)*s1hat*erfcinv(2*t))/s2hat^2 vROC=t(c(dm1,ds1,dm2,ds2))%*% S %*% t(t(c(dm1,ds1,dm2,ds2))) CIinvROC=c(norminvROC-qnorm(1-alpha/2)*sqrt(vROC), norminvROC+qnorm(1-alpha/2)*sqrt(vROC)) CIROC=c( pnorm(CIinvROC[1]), pnorm(CIinvROC[2])) CIROCvec1[kk]=CIROC[1] CIROCvec2[kk]=CIROC[2] Sehat[kk]=pnorm(norminvROC) #points(t,Sehat[length(Sehat)], col = "red") #lines(c(t,t),c(CIROC[1],CIROC[1]),col="black",lty=2) #lines(c(t,t),c(CIROC[2],CIROC[2]),col="black") kk=kk+1 } if (plots=="on"){ lines(c(tt),c(CIROCvec1),col="black",lty=2) lines(c(tt),c(CIROCvec2),col="black",lty=2) #====END GRID LINES========================= #============================================= # x11() # plot.new() txt <- paste("Box-Cox Based ROC, alpha =", round(alpha,3) ) plot(linspace(0,1,1000),roc(linspace(0,1,1000),transx,transy),main=" ",xlab="FPR",ylab="TPR",type="l",col="red") lines(linspace(0,1,10),linspace(0,1,10),type="l", lty=2) title(main=txt) } kk=1;CIROC=c(1,1);CIROCvec1=0;CIROCvec2=0;Sehat=0; tt=c(givenSP);tt=1-givenSP; #tt=c(0.05,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9) for (t in tt){ norminvROC= ((m2hat-m1hat)/s2hat+s1hat/s2hat*qnorm(t)) #====================ROC1============================== dm1 =-1/s2hat; ds1 =-(2^(1/2)*erfcinv(2*t))/s2hat; dm2 =1/s2hat; ds2 =(m1hat - m2hat)/s2hat^2 + (2^(1/2)*s1hat*erfcinv(2*t))/s2hat^2 vROC=t(c(dm1,ds1,dm2,ds2))%*% S %*% t(t(c(dm1,ds1,dm2,ds2))) CIinvROC=c(norminvROC-qnorm(1-alpha/2)*sqrt(vROC), norminvROC+qnorm(1-alpha/2)*sqrt(vROC)) CIROC=c( pnorm(CIinvROC[1]), pnorm(CIinvROC[2])) CIROCvec1[kk]=CIROC[1] CIROCvec2[kk]=CIROC[2] Sehat[kk]=pnorm(norminvROC) if (plots=="on"){ points(t,Sehat[length(Sehat)], col = "green") lines(c(t,t),c(CIROC[1],CIROC[2]),col="green") } kk=kk+1 } Spvalues=1-tt;FPRvalues=tt; SEandCIs = matrix( c(Spvalues, Sehat, CIROCvec1,CIROCvec2), nrow=length(CIROCvec1), ncol=4) colnames(SEandCIs) <- c("Given Sp", "Sehat","LL of 95%CI","UL of 95%CI") SEandCIs CIlowSe=CIROCvec1 CIuppSe=CIROCvec2 CIse=t(rbind(CIlowSe,CIuppSe)) #==================ROCinv1 (CIs for Sp)=================== mm1=m1hat; mm2=m2hat; ss1=s1hat; ss2=s2hat; m1hat=mm2; m2hat=mm1; s1hat=ss2; s2hat=ss1; kk=1;CIROC=c(1,1);CIROCvec1=0;CIROCvec2=0;Sphat=0; mh=m1hat; md=m2hat; n=n1; m=n2; xlam=transx; ylam=transy; lam=cc$transformation.parameter I[1,1]=n/sh^2; I[2,2]=-(n/sh^2-3/sh^4*sum((xlam-mh)^2)); I[3,3]=m/sd^2; I[4,4]=-(m/sd^2-3/sd^4*sum((ylam-md)^2)); kk= sum(((mh - (x^lam - 1)/lam)*((2*(x^lam - 1))/lam^3 - (2*x^lam*log(x))/lam^2 + (x^lam*log(x)^2)/lam))/sh^2) + - sum(((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)^2/sd^2) + - sum(((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)^2/sh^2) + + sum(((md - (y^lam - 1)/lam)*((2*(y^lam - 1))/lam^3 - (2*y^lam*log(y))/lam^2 + (y^lam*log(y)^2)/lam))/sd^2); I[5,5]=-kk ; I[1,5]=sum(((2*(x^lam - 1))/lam^2 - (2*x^lam*log(x))/lam)/(2*sh^2)); I[2,5]=-sum((2*((x^lam - 1)/lam^2 - (x^lam*log(x))/lam)*(mh - (x^lam - 1)/lam))/sh^3); I[3,5]=-sum(-((2*(y^lam - 1))/lam^2 - (2*y^lam*log(y))/lam)/(2*sd^2)); I[4,5]=-sum((2*((y^lam - 1)/lam^2 - (y^lam*log(y))/lam)*(md - (y^lam - 1)/lam))/sd^3); I[5,1]=I[1,5] I[5,2]=I[2,5] I[5,3]=I[3,5] I[5,4]=I[4,5] S=inv(I) S=S[1:4,1:4] pp=c(givenSE); #pp=c(0.05,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9) kk=1;CIROC=c(1,1);CIROCvec1=0;CIROCvec2=0;Sphat=0; for (t in pp){ norminvROC= ((m2hat-m1hat)/s2hat+s1hat/s2hat*qnorm(t)) #====================ROC1============================== dm1 =-1/s2hat; ds1 =-(2^(1/2)*erfcinv(2*t))/s2hat; dm2 =1/s2hat; ds2 =(m1hat - m2hat)/s2hat^2 + (2^(1/2)*s1hat*erfcinv(2*t))/s2hat^2 vROC=t(c(dm1,ds1,dm2,ds2))%*% S %*% t(t(c(dm1,ds1,dm2,ds2))) CIinvROC=c(norminvROC-qnorm(1-alpha/2)*sqrt(vROC), norminvROC+qnorm(1-alpha/2)*sqrt(vROC)) CIROC=c( pnorm(CIinvROC[1]), pnorm(CIinvROC[2])) CIROCvec1[kk]=CIROC[1] CIROCvec2[kk]=CIROC[2] Sphat[kk]=pnorm(norminvROC) if (plots=="on"){ points(Sphat[length(Sphat)],t, col = "purple") lines(c(CIROC[1],CIROC[2]),c(t,t),col="purple") } kk=kk+1 } #SPandCIs = matrix( c(Sphat, CIROCvec1,CIROCvec2), nrow=length(CIROCvec1), ncol=3) #colnames(SPandCIs) <- c("Sphat","LL of 95%CI","UL of 95%CI") #SPandCIs Sevalues=givenSE; SPandCIs = matrix( c(Sevalues, Sphat, CIROCvec1,CIROCvec2), nrow=length(CIROCvec1), ncol=4) colnames(SPandCIs) <- c("Given Se", "Sphat","LL of 95%CI","UL of 95%CI") SPandCIs CIlowSp=CIROCvec1 CIuppSp=CIROCvec2 CIsp=t(rbind(CIlowSp,CIuppSp)) return(list(SPandCIs=formattable(as.matrix(SPandCIs), digits = 4, format = "f"), SEandCIs=formattable(as.matrix(SEandCIs), digits = 4, format = "f"), Sevalues=Sevalues, Sphat=Sphat, CIsp=formattable(as.matrix(CIsp), digits = 4, format = "f"), Spvalues=Spvalues, Sehat=Sehat, CIse=formattable(as.matrix(CIse), digits = 4, format = "f"))) } } ``` ```{r, echo=FALSE, eval=TRUE} checkboxcox<-function(marker, D, plots, printShapiro = FALSE){ oldpar <- par(no.readonly = TRUE) on.exit(par(oldpar)) if (plots!="on"){plots="off"} if (length(marker) != length(D)) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (sum(is.na(marker)) > 0 | sum(is.na(D)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if (sum(marker < 0) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { statusD=D; xor=marker[D==0] yor=marker[D==1] x = xor y = yor scores = marker # scores = c(x, y) #====THIS IS THE LIKELIHOOD OF THE BOXCOX TRANSFORMATION======================================= #INPUT ARGUMENTS: TWO GROUPS (x,y) and plots="on" or "off"==================================== likbox<-function(x,y,h){ n=length(x); m=length(y); out=c(); for (i in 1:length(h)){ #print(i) if (h[i]==0){ xh=log(x); yh=log(y); } else { xh=((x^h[i])-1)/h[i]; yh=((y^h[i])-1)/h[i]; } out[i]<-c( -n/2*log(sum((xh-sum(xh)/n)^2)/n) -m/2*log(sum((yh-sum(yh)/m)^2)/m) +(h[i]-1)*(sum(log(x))+sum(log(y)))) } return(out) } #====THIS THE FUNCTION NEEDED TO APPLY THE BOXCOX TRANSFORMATION INPUT ARGUMENTS: #====TWO GROUPS (x,y) and plots="on" or "off"==================================== boxcoxleo<-function(x,y,plots){ if (plots!="on"){plots="off"} init=1; logL<-function(h){ -likbox(x,y,h) } #lam=fminsearch(logL,init) # lam <- optimize(logL, c(-100, 100), tol = 0.0001) # lam=c(lam$minimum) #c(lam$optbase$xopt) lam<- optim(1,logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) transx=((x^lam)-1)/lam transy=((y^lam)-1)/lam if (!printShapiro) { test1=shapiro.test(x) test2=shapiro.test(y) test3=shapiro.test(transx) test4=shapiro.test(transy) } if (printShapiro) { test1=print(shapiro.test(x)) test2=print(shapiro.test(y)) test3=print(shapiro.test(transx)) test4=print(shapiro.test(transy)) } pval1=test1$p.value pval2=test2$p.value pval3=test3$p.value pval4=test4$p.value pvalues=c(pval1,pval2,pval3,pval4) return(list(transformation.parameter=lam,transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam,pvalues=pvalues)) } #====APPLY THE BOXCOX TRANSFORMATION WITH OR WITHOUT THE PLOTS================ cc=boxcoxleo(x,y,plots) lam=cc$transformation.parameter pvalues=cc$pvalues transx=cc$transx #transformed scores for healthy transy=cc$transy #transformed scores for diseased if (plots!="on"){plots="off"} #====IF PLOTS ARE REQUESTED PROVIDE THE HISTOGRAMS AND QQ-PLOTS FOR EACH GROUP #====BEFORE AND AFTER THE TRANSFORMATION======================================= if (plots=="on") { # x11() par(mfrow=c(2,2)) par(mar = c(4.1, 3.1, 3.1, 1.1)) hist(x, main = "Histogram of X", xlab = "X") qqnorm(x, main = "Normal Q-Q Plot of X") qqline(x, col = 2,lwd=2,lty=2) hist(transx, main = "Histogram of Transformed X", xlab = "Transformed X") qqnorm(transx, main = "Normal Q-Q Plot of Transformed X") qqline(transx, col = 2,lwd=2,lty=2) par(mfrow=c(2,2)) hist(y, main = "Histogram of Y", xlab = "Y") qqnorm(y, main = "Normal Q-Q Plot of Y") qqline(y, col = 2,lwd=2,lty=2) hist(transy, main = "Histogram of Transformed Y", xlab = "Transformed Y") qqnorm(transy, main = "Normal Q-Q Plot of Transformed Y") qqline(transy, col = 2,lwd=2,lty=2) par(mar = c(5.1, 4.1, 4.1, 2.1)) par(mfrow=c(1,1)) } #====TWO ROC FUNCTIONS: ONE IS THE BOXCOX AND THE OTHER THE REFERENCE LINE================ roc<-function(t){ na = length(transx) nb = length(transy) stransx=sqrt( var(transx)*(na-1)/na ) stransy=sqrt( var(transy)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(transx), sd=stransx), mean=mean(transy), sd=stransy) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } #================IF PLOTS ARE REQUESTED PLOT THE ROCS== if (plots=="on") { # x11() par(mfrow = c(1, 1)) plot(linspace(0,1,1000),roc(linspace(0,1,1000)),main="Empirical and Box-Cox Based ROC",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="red") lines(linspace(0,1,1000),linspace(0,1,1000),type="l", lty=2) lines(roc.curve(scores, statusD), col=1) #================LEGEND OF THE ROC PLOT================ legend("bottomright", legend=c(paste("Box-Cox ROC estimate "), paste("Empirical ROC estimate ")), col=c("red", "black"), lty=c(1, 1), pch = c(NA, NA), cex=0.8) } #================OUTPUT ARGUMENTS====================== return(list(transformation.parameter=lam, transx=((x^lam)-1)/lam, transy=((y^lam)-1)/lam, pval_x=pvalues[1], pval_y=pvalues[2], pval_transx=pvalues[3], pval_transy=pvalues[4], rocfun=roc )) } } ``` ```{r, echo=FALSE, eval=TRUE} threerocs=function(marker, D, plots) { x=marker[D==0]; y=marker[D==1]; # Plot the empirical ROC curve roc_obj <- roc(D, marker, quiet = TRUE) auc_value <- auc(roc_obj) tpr=roc_obj$sensitivities fpr=1-roc_obj$specificities if (plots == "on") { plot(fpr, tpr, type = "l", col = "blue", xlim = c(0, 1), ylim = c(0, 1), xlab = "FPR = 1 - Specificity", ylab = "TPR = Sensitivity", main = "Three Estimates of the ROC Curve") } t=linspace(0,1,10000) #set a grid for the FPR ############## BOX-COX ############################# roctbc=checkboxcox(marker,D, plots="off", printShapiro = FALSE); myroc=rocboxcox(marker,D,0.05,plots="off",FALSE) rocboxc=myroc$rocbc AUCbc=myroc$AUC if (plots == "on") {lines(t,roctbc$roc(t),col="red")} ############## Metz ############################# curve2= roc_curves(D,marker,method="binormal") rocmrm=points(curve2, metric="specificity", values=1-c(t)) rocmrm$TPR AUCmrm=trapz(t,rocmrm$TPR) if (plots == "on") {lines(t,rocmrm$TPR,col="forestgreen", lty = "dashed")} #legend(x = "right", legend=c("one","two","three")) if (plots == "on") { legend("bottomright", legend = c(paste("Empirical (AUC = ", formattable(auc_value, digits = 4, format = "f"), ")", sep = ""), paste("Box-Cox (AUC = ", formattable(AUCbc, digits = 4, format = "f"), ")", sep = ""), paste("Metz (AUC = ", formattable(AUCmrm, digits = 4, format = "f"), ")", sep = "")), col = c("blue", "red", "forestgreen"), lty = c("solid", "dotted", "dashed"), cex = 0.8) } return(list(AUC_Empirical = as.numeric(auc_value), AUC_Metz = AUCmrm, AUC_BoxCox = AUCbc)) } ``` ```{r, echo=FALSE, eval=TRUE} checkboxcox2 <-function (marker1, marker2, D, plots, printShapiro = FALSE){ if ((length(marker1) != length(D)) | (length(marker2) != length(D))) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if ((sum(marker1 < 0) > 0) | (sum(marker2 < 2)) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { erf <- function (x) 2 * pnorm(x * sqrt(2)) - 1 erfc <- function (x) 2 * pnorm(x * sqrt(2), lower.tail = FALSE) erfinv <- function (x) qnorm((1 + x)/2)/sqrt(2) erfcinv <- function (x) qnorm(x/2, lower.tail = FALSE)/sqrt(2) if (plots!="on"){plots="off"} W1a=marker1[D==0] W1b=marker1[D==1] W2a=marker2[D==0] W2b=marker2[D==1] Wa=cbind(W1a,W2a); Wb=cbind(W1b,W2b); na=length(W1a) nb=length(W1b) likbox2D<-function(W1a,W1b,W2a,W2b,lam){ na=length(W1a); nb=length(W1b); out=c(); #for (i in 1:length(h)){ W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) cova= cov(Walam)*(na-1)/na; covb= cov(Wblam)*(nb-1)/nb; mualam=c(mean(W1alam), mean(W2alam)) mublam=c(mean(W1blam), mean(W2blam)) out= -sum(log(dmvnorm(Walam,mualam,cova)))-sum(log(dmvnorm(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b)); return(out) } lam=c(2,2) likbox2D(W1a,W1b,W2a,W2b,lam) logL<-function(h){ likbox2D(W1a,W1b,W2a,W2b,h) } # init=c(0.8,0.8) # lam=fminsearch(logL,init) # lam=c(lam$optbase$xopt) lam<- optim(c(1,1),logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) lam out <- optim(c(1,1), logL, method = "Nelder-Mead") lam = out$par ################################### W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) m1ahat=mean(W1alam) m1bhat=mean(W1blam) m2ahat=mean(W2alam) m2bhat=mean(W2blam) s1ahat=sqrt( var(W1alam)*(na-1)/na ) s1bhat=sqrt( var(W1blam)*(nb-1)/nb ) s2ahat=sqrt( var(W2alam)*(na-1)/na ) s2bhat=sqrt( var(W2blam)*(nb-1)/nb ) cova= cov(Walam)*(na-1)/na;cova=cova[1,2]; covb= cov(Wblam)*(nb-1)/nb;covb=covb[1,2]; lam1=lam[1] lam2=lam[2] # ==== PERFORM SHAPIRO-WILK TESTS ============================================================= x1 = W1a; y1 = W1b; transx1 = W1alam; transy1 = W1blam x2 = W2a; y2 = W2b; transx2 = W2alam; transy2 = W2blam if (!printShapiro) { test_x1=shapiro.test(x1) test_y1=shapiro.test(y1) test_t_x1=shapiro.test(transx1) test_t_y1=shapiro.test(transy1) test_x2=shapiro.test(x2) test_y2=shapiro.test(y2) test_t_x2=shapiro.test(transx2) test_t_y2=shapiro.test(transy2) } if (printShapiro) { test_x1=print(shapiro.test(x1)) test_y1=print(shapiro.test(y1)) test_t_x1=print(shapiro.test(transx1)) test_t_y1=print(shapiro.test(transy1)) test_x2=print(shapiro.test(x2)) test_y2=print(shapiro.test(y2)) test_t_x2=print(shapiro.test(transx2)) test_t_y2=print(shapiro.test(transy2)) } pval_x1=test_x1$p.value pval_y1=test_y1$p.value pval_t_x1=test_t_x1$p.value pval_t_y1=test_t_y1$p.value pval_x2=test_x2$p.value pval_y2=test_y2$p.value pval_t_x2=test_t_x2$p.value pval_t_y2=test_t_y2$p.value pvalues=c(pval_x1, pval_y1, pval_t_x1, pval_t_y1, pval_x2, pval_y2, pval_t_x2, pval_t_y2) res_shapiro = list(test_x1 = test_x1, test_y1 = test_y1, test_t_x1 = test_t_x1, test_t_y1 = test_t_y1, test_x2 = test_x2, test_y2 = test_y2, test_t_x2 = test_t_x2, test_t_y2 = test_t_y2) #====TWO ROC FUNCTIONS: ONE IS THE BOXCOX AND THE OTHER THE REFERENCE LINE================ roc1<-function(t){ na = length(W1alam) nb = length(W1blam) s1alam=sqrt( var(W1alam)*(na-1)/na ) s1blam=sqrt( var(W1blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W1alam), sd=s1alam), mean=mean(W1blam), sd=s1blam) } roc2<-function(t){ na = length(W2alam) nb = length(W2blam) s2alam=sqrt( var(W2alam)*(na-1)/na ) s2blam=sqrt( var(W2blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W2alam), sd=s2alam), mean=mean(W2blam), sd=s2blam) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } #================IF PLOTS ARE REQUESTED PLOT ALL OF THE THINGS==== if (plots=="on") { par(mfrow=c(2,2)) par(mar = c(4.1, 3.1, 3.1, 1.1)) # HISTOGRAMS & QQ, X1 hist(W1a, main = "Histogram of X1", xlab = "X1") qqnorm(W1a, main = "Normal Q-Q Plot of X1") qqline(W1a, col = 2,lwd=2,lty=2) hist(W1alam, main = "Histogram of Transformed X1", xlab = "Transformed X1") qqnorm(W1alam, main = "Normal Q-Q Plot of Transformed X1") qqline(W1alam, col = 2,lwd=2,lty=2) # HISTOGRAMS & QQ, Y1 hist(W1b, main = "Histogram of Y1", xlab = "Y1") qqnorm(W1b, main = "Normal Q-Q Plot of Y1") qqline(W1b, col = 2,lwd=2,lty=2) hist(W1blam, main = "Histogram of Transformed Y1", xlab = "Transformed Y1") qqnorm(W1blam, main = "Normal Q-Q Plot of Transformed Y1") qqline(W1blam, col = 2,lwd=2,lty=2) # HISTOGRAMS & QQ, X2 hist(W2a, main = "Histogram of X2", xlab = "X2") qqnorm(W2a, main = "Normal Q-Q Plot of X2") qqline(W2a, col = 2,lwd=2,lty=2) hist(W2alam, main = "Histogram of Transformed X2", xlab = "Transformed X2") qqnorm(W2alam, main = "Normal Q-Q Plot of Transformed X2") qqline(W2alam, col = 2,lwd=2,lty=2) # HISTOGRAMS & QQ, Y2 hist(W2b, main = "Histogram of Y2", xlab = "Y2") qqnorm(W2b, main = "Normal Q-Q Plot of Y2") qqline(W2b, col = 2,lwd=2,lty=2) hist(W2blam, main = "Histogram of Transformed Y2", xlab = "Transformed Y2") qqnorm(W2blam, main = "Normal Q-Q Plot of Transformed Y2") qqline(W2blam, col = 2,lwd=2,lty=2) par(mar = c(5.1, 4.1, 4.1, 2.1)) # ROC PLOT 1 par(mfrow = c(1, 2)) plot(linspace(0,1,1000), roc1(linspace(0,1,1000)), main="Marker1", xlab="FPR = 1 - Specificity", ylab="TPR = Sensitivity", type="l", col="red") lines(linspace(0,1,1000), linspace(0,1,1000), type="l", lty=2) lines(roc.curve(marker1, D), col="red", lty = 3, lwd = 1.5) # LEGEND legend("bottomright", legend=c(paste("Box-Cox ROC Estimate"), paste("Empirical ROC Estimate")), col=c("red", "red"), lty=c(1, 3), lwd = c(1, 1.5), pch = c(NA, NA), cex=0.8) # ROC PLOT 2 plot(linspace(0,1,1000), roc2(linspace(0,1,1000)), main="Marker2", xlab="FPR = 1 - Specificity", ylab="TPR = Sensitivity", type="l", col="black") lines(linspace(0,1,1000), linspace(0,1,1000), type="l", lty=2) lines(roc.curve(marker2, D), col="black", lty = 3, lwd = 1.5) # LEGEND legend("bottomright", legend=c(paste("Box-Cox ROC estimate"), paste("Empirical ROC estimate")), col=c("black", "black"), lty=c(1, 3), lwd = c(1, 1.5), pch = c(NA, NA), cex=0.8) par(mfrow = c(1, 1)) } return(list(res_shapiro = res_shapiro, transformation.parameter.1 = lam[1], transx1 = W1alam, transy1 = W1blam, transformation.parameter.2 = lam[2], transx2 = W2alam, transy2 = W2blam, pval_x1 = pvalues[1], pval_y1 = pvalues[2], pval_transx1 = pvalues[3], pval_transy1 = pvalues[4], pval_x2 = pvalues[5], pval_y2 = pvalues[6], pval_transx2 = pvalues[7], pval_transy2 = pvalues[8], roc1 = roc1, roc2 = roc2)) } } ``` ```{r, echo=FALSE, eval=TRUE} comparebcAUC <-function (marker1, marker2, D, alpha, plots){ if ((length(marker1) != length(D)) | (length(marker2) != length(D))) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (alpha <= 0 | alpha >= 1) { stop("ERROR: The level of significance, alpha, should be set between 0 and 1. A common choice is 0.05.") } else if (sum(is.na(marker1)) > 0 | sum(is.na(marker2)) > 0 | sum(is.na(D)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if ((sum(marker1 < 0) > 0) | (sum(marker2 < 2)) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { erf <- function (x) 2 * pnorm(x * sqrt(2)) - 1 erfc <- function (x) 2 * pnorm(x * sqrt(2), lower.tail = FALSE) erfinv <- function (x) qnorm((1 + x)/2)/sqrt(2) erfcinv <- function (x) qnorm(x/2, lower.tail = FALSE)/sqrt(2) pmalpha=qnorm(1-alpha/2); if (plots!="on"){plots="off"} W1a=marker1[D==0] W1b=marker1[D==1] W2a=marker2[D==0] W2b=marker2[D==1] Wa=cbind(W1a,W2a); Wb=cbind(W1b,W2b); na=length(W1a) nb=length(W1b) likbox2D<-function(W1a,W1b,W2a,W2b,lam){ na=length(W1a); nb=length(W1b); out=c(); #for (i in 1:length(h)){ W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) #cova=1/na*sum((W1alam-mean(W1alam))*(W2alam-mean(W2alam))) #covb=1/nb*sum((W1blam-mean(W1blam))*(W2blam-mean(W2blam))) cova= cov(Walam)*(na-1)/na; covb= cov(Wblam)*(nb-1)/nb; mualam=c(mean(W1alam), mean(W2alam)) mublam=c(mean(W1blam), mean(W2blam)) #dmvn(c(0,0), mu, mcov, log = F) # out= (-sum(log(dmvn(Walam,mualam,cova)))-sum(log(dmvn(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b))); out= -sum(log(dmvnorm(Walam,mualam,cova)))-sum(log(dmvnorm(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b)); #out1=-sum(log(dmvnorm(Walam,mualam,cova))) #out2=-sum(log(dmvnorm(Wblam,mublam,covb))) #out3=-(lam[1]-1)*sum(log(W1a)) #out4=-(lam[2]-1)*sum(log(W2a)) #out5=-(lam[1]-1)*sum(log(W1b)) #out6=-(lam[2]-1)*sum(log(W2b)) #out=c(out,out1,out2,out3,out4,out5,out6, cova, covb) #out= (-sum(log(mvnpdf2([W1alam(g(1)) W2alam(g(2))],[mean(W1alam(g(1))) mean(W2alam(g(2)))],[std(W1alam(g(1)),1).^2 cova;cova (std(W2alam(g(2)),1)).^2])))+... # -sum(log(mvnpdf2([W1blam(g(1)) W2blam(g(2))],[mean(W1blam(g(1))) mean(W2blam(g(2)))],[std(W1blam(g(1)),1).^2 covb;covb (std(W2blam(g(2)),1)).^2])))+... # -(g(1)-1).*sum(log(W1a))-(g(2)-1).*sum(log(W2a))-(g(1)-1).*sum(log(W1b))-(g(2)-1).*sum(log(W2b))); return(out) } lam=c(2,2) likbox2D(W1a,W1b,W2a,W2b,lam) logL<-function(h){ likbox2D(W1a,W1b,W2a,W2b,h) } # init=c(0.8,0.8) # lam=fminsearch(logL,init) # lam=c(lam$optbase$xopt) lam<- optim(c(1,1),logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) lam #init=c(-10,10) #lam=optimize(logL,init) #lam #f <- function (x, a) (x - a)^2 #xmin <- optimize(f, c(0, 1), tol = 0.0001, a = 1/3) #xmin out <- optim(c(1,1), logL, method = "Nelder-Mead") lam = out$par ################################### W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) m1ahat=mean(W1alam) m1bhat=mean(W1blam) m2ahat=mean(W2alam) m2bhat=mean(W2blam) s1ahat=sqrt( var(W1alam)*(na-1)/na ) s1bhat=sqrt( var(W1blam)*(nb-1)/nb ) s2ahat=sqrt( var(W2alam)*(na-1)/na ) s2bhat=sqrt( var(W2blam)*(nb-1)/nb ) cova= cov(Walam)*(na-1)/na;cova=cova[1,2]; covb= cov(Wblam)*(nb-1)/nb;covb=covb[1,2]; lam1=lam[1] lam2=lam[2] ################################## h1_1=sum(-2/(2*s1ahat^2 - (2*cova^2)/s2ahat^2)) h1_2=sum(-(2*s1ahat*s2ahat^2*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_3=0 h1_4=0; h1_5=sum(cova/(- cova^2 + s1ahat^2*s2ahat^2)); h1_6=sum((s2ahat*(2*cova^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat) - 2*cova*s1ahat^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_7=0 h1_8=0 h1_9=sum((cova^2*(lam2*m2ahat - W2a^lam2 + 1) + s2ahat^2*((lam2*m2ahat - W2a^lam2 + 1)*s1ahat^2 - 2*cova*lam2*m1ahat))/(lam2*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s2ahat^2*(2*lam2 - 2*W1a^lam1*lam2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_10=0 h1_11=sum((2*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h1_12=sum(-(cova*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- cova^2 + s1ahat^2*s2ahat^2))) h2_1=sum(-(2*s1ahat*s2ahat^2*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_2=sum(- (3*s1ahat^2*s2ahat^6 + cova^4*(lam1^2*m2ahat^2 + lam1^2*s2ahat^2) - cova^3*(2*lam1*m2ahat*s2ahat^2 - 2*W1a^lam1*lam1*m2ahat*s2ahat^2 + 2*lam1^2*m1ahat*m2ahat*s2ahat^2) - cova*(6*lam1*m2ahat*s1ahat^2*s2ahat^4 - 6*W1a^lam1*lam1*m2ahat*s1ahat^2*s2ahat^4 + 6*lam1^2*m1ahat*m2ahat*s1ahat^2*s2ahat^4) + cova^2*(W1a^(2*lam1)*s2ahat^4 - 2*W1a^lam1*s2ahat^4 + s2ahat^4 + 2*lam1*m1ahat*s2ahat^4 + lam1^2*m1ahat^2*s2ahat^4 - 2*W1a^lam1*lam1*m1ahat*s2ahat^4 + 3*lam1^2*m2ahat^2*s1ahat^2*s2ahat^2) - 6*W1a^lam1*s1ahat^2*s2ahat^6 + 3*W1a^(2*lam1)*s1ahat^2*s2ahat^6 - lam1^2*s1ahat^4*s2ahat^6 + 3*lam1^2*m1ahat^2*s1ahat^2*s2ahat^6 + 6*lam1*m1ahat*s1ahat^2*s2ahat^6 - 6*W1a^lam1*lam1*m1ahat*s1ahat^2*s2ahat^6)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^4*(W2a^(2*lam2)*lam1^2 - 2*W2a^lam2*lam1^2 + lam1^2) + cova^2*(3*lam1^2*s1ahat^2*s2ahat^2 + 3*W2a^(2*lam2)*lam1^2*s1ahat^2*s2ahat^2 - 6*W2a^lam2*lam1^2*s1ahat^2*s2ahat^2) - lam2*((2*W2a^lam2*lam1^2*m2ahat - 2*lam1^2*m2ahat)*cova^4 + (2*lam1*s2ahat^2 + 2*lam1^2*m1ahat*s2ahat^2 - 2*W1a^lam1*lam1*s2ahat^2 - 2*W2a^lam2*lam1*s2ahat^2 - 2*W2a^lam2*lam1^2*m1ahat*s2ahat^2 + 2*W1a^lam1*W2a^lam2*lam1*s2ahat^2)*cova^3 + (6*W2a^lam2*lam1^2*m2ahat*s1ahat^2*s2ahat^2 - 6*lam1^2*m2ahat*s1ahat^2*s2ahat^2)*cova^2 + (6*lam1*s1ahat^2*s2ahat^4 + 6*lam1^2*m1ahat*s1ahat^2*s2ahat^4 - 6*W1a^lam1*lam1*s1ahat^2*s2ahat^4 - 6*W2a^lam2*lam1*s1ahat^2*s2ahat^4 - 6*W2a^lam2*lam1^2*m1ahat*s1ahat^2*s2ahat^4 + 6*W1a^lam1*W2a^lam2*lam1*s1ahat^2*s2ahat^4)*cova))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_3=sum(0) h2_4=sum(0) h2_5=sum((s1ahat*(2*cova^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat) - 2*cova*s2ahat^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_6=sum((lam2*((2*cova*s2ahat^3*(2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat) - 2*cova*s2ahat*(4*cova*lam1^2*m2ahat - 4*W2a^lam2*cova*lam1^2*m2ahat))*s1ahat^3 + 2*cova*s2ahat*(2*cova^2*lam1 + 2*cova^2*lam1^2*m1ahat - 2*W1a^lam1*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1^2*m1ahat + 2*W1a^lam1*W2a^lam2*cova^2*lam1)*s1ahat) - 2*cova*s1ahat^3*s2ahat*(2*cova*lam1^2 + 2*W2a^(2*lam2)*cova*lam1^2 - 4*W2a^lam2*cova*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - ((4*cova^2*lam1^2*m2ahat^2*s2ahat - 2*cova*s2ahat^3*(2*lam1*m2ahat + cova*lam1^2 + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat))*s1ahat^3 + (2*cova*(2*cova + 2*W1a^(2*lam1)*cova - 4*W1a^lam1*cova + 2*cova*lam1^2*m1ahat^2 + 4*cova*lam1*m1ahat - 4*W1a^lam1*cova*lam1*m1ahat)*s2ahat^3 + 2*cova*(cova^3*lam1^2 - 2*cova^2*lam1*m2ahat + 2*W1a^lam1*cova^2*lam1*m2ahat - 2*cova^2*lam1^2*m1ahat*m2ahat)*s2ahat)*s1ahat)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_7=sum(0) h2_8=sum(0) h2_9=sum(- (cova*(s2ahat^4*(- 4*lam2^2*m1ahat^2*s1ahat + 2*lam2^2*s1ahat^3) - s1ahat^3*s2ahat^2*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2)) - cova^3*(s1ahat*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2) + 2*lam2^2*s1ahat*s2ahat^2) + s1ahat^3*s2ahat^4*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + cova^2*s1ahat*s2ahat^2*(6*lam2*m1ahat + 6*lam2^2*m1ahat*m2ahat - 6*W2a^lam2*lam2*m1ahat))/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2ahat - 6*W1a^lam1*lam2 - 6*W2a^lam2*lam2 + 6*W1a^lam1*W2a^lam2*lam2 - 6*W1a^lam1*lam2^2*m2ahat)*cova^2*s1ahat*s2ahat^2 + (8*W1a^lam1*lam2^2*m1ahat - 8*lam2^2*m1ahat)*cova*s1ahat*s2ahat^4 + (2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat)*s1ahat^3*s2ahat^4) - cova*s1ahat*s2ahat^4*(4*W1a^(2*lam1)*lam2^2 - 8*W1a^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_10=sum(0) h2_11=sum((2*s1ahat*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_12=sum(-(2*cova*s1ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h3_1=sum(0) h3_2=sum(0) h3_3=sum(-2/(2*s1bhat^2 - (2*covb^2)/s2bhat^2)) h3_4=sum(-(2*s1bhat*s2bhat^2*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_5=sum(0) h3_6=sum(0) h3_7=sum(covb/(- covb^2 + s1bhat^2*s2bhat^2)) h3_8=sum((s2bhat*(2*covb^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat) - 2*covb*s1bhat^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_9=sum(0) h3_10=sum((covb^2*(lam2*m2bhat - W2b^lam2 + 1) + s2bhat^2*((lam2*m2bhat - W2b^lam2 + 1)*s1bhat^2 - 2*covb*lam2*m1bhat))/(lam2*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s2bhat^2*(2*lam2 - 2*W1b^lam1*lam2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_11=sum((2*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h3_12=sum(-(covb*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- covb^2 + s1bhat^2*s2bhat^2))) h4_1=sum(0) h4_2=sum(0) h4_3=sum(-(2*s1bhat*s2bhat^2*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_4=sum(- (3*s1bhat^2*s2bhat^6 + covb^4*(lam1^2*m2bhat^2 + lam1^2*s2bhat^2) - covb^3*(2*lam1*m2bhat*s2bhat^2 - 2*W1b^lam1*lam1*m2bhat*s2bhat^2 + 2*lam1^2*m1bhat*m2bhat*s2bhat^2) - covb*(6*lam1*m2bhat*s1bhat^2*s2bhat^4 - 6*W1b^lam1*lam1*m2bhat*s1bhat^2*s2bhat^4 + 6*lam1^2*m1bhat*m2bhat*s1bhat^2*s2bhat^4) + covb^2*(W1b^(2*lam1)*s2bhat^4 - 2*W1b^lam1*s2bhat^4 + s2bhat^4 + 2*lam1*m1bhat*s2bhat^4 + lam1^2*m1bhat^2*s2bhat^4 - 2*W1b^lam1*lam1*m1bhat*s2bhat^4 + 3*lam1^2*m2bhat^2*s1bhat^2*s2bhat^2) - 6*W1b^lam1*s1bhat^2*s2bhat^6 + 3*W1b^(2*lam1)*s1bhat^2*s2bhat^6 - lam1^2*s1bhat^4*s2bhat^6 + 3*lam1^2*m1bhat^2*s1bhat^2*s2bhat^6 + 6*lam1*m1bhat*s1bhat^2*s2bhat^6 - 6*W1b^lam1*lam1*m1bhat*s1bhat^2*s2bhat^6)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^4*(W2b^(2*lam2)*lam1^2 - 2*W2b^lam2*lam1^2 + lam1^2) + covb^2*(3*lam1^2*s1bhat^2*s2bhat^2 + 3*W2b^(2*lam2)*lam1^2*s1bhat^2*s2bhat^2 - 6*W2b^lam2*lam1^2*s1bhat^2*s2bhat^2) - lam2*((2*W2b^lam2*lam1^2*m2bhat - 2*lam1^2*m2bhat)*covb^4 + (2*lam1*s2bhat^2 + 2*lam1^2*m1bhat*s2bhat^2 - 2*W1b^lam1*lam1*s2bhat^2 - 2*W2b^lam2*lam1*s2bhat^2 - 2*W2b^lam2*lam1^2*m1bhat*s2bhat^2 + 2*W1b^lam1*W2b^lam2*lam1*s2bhat^2)*covb^3 + (6*W2b^lam2*lam1^2*m2bhat*s1bhat^2*s2bhat^2 - 6*lam1^2*m2bhat*s1bhat^2*s2bhat^2)*covb^2 + (6*lam1*s1bhat^2*s2bhat^4 + 6*lam1^2*m1bhat*s1bhat^2*s2bhat^4 - 6*W1b^lam1*lam1*s1bhat^2*s2bhat^4 - 6*W2b^lam2*lam1*s1bhat^2*s2bhat^4 - 6*W2b^lam2*lam1^2*m1bhat*s1bhat^2*s2bhat^4 + 6*W1b^lam1*W2b^lam2*lam1*s1bhat^2*s2bhat^4)*covb))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_5=sum(0) h4_6=sum(0) h4_7=sum((s1bhat*(2*covb^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat) - 2*covb*s2bhat^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_8=sum((lam2*((2*covb*s2bhat^3*(2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat) - 2*covb*s2bhat*(4*covb*lam1^2*m2bhat - 4*W2b^lam2*covb*lam1^2*m2bhat))*s1bhat^3 + 2*covb*s2bhat*(2*covb^2*lam1 + 2*covb^2*lam1^2*m1bhat - 2*W1b^lam1*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1^2*m1bhat + 2*W1b^lam1*W2b^lam2*covb^2*lam1)*s1bhat) - 2*covb*s1bhat^3*s2bhat*(2*covb*lam1^2 + 2*W2b^(2*lam2)*covb*lam1^2 - 4*W2b^lam2*covb*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - ((4*covb^2*lam1^2*m2bhat^2*s2bhat - 2*covb*s2bhat^3*(2*lam1*m2bhat + covb*lam1^2 + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat))*s1bhat^3 + (2*covb*(2*covb + 2*W1b^(2*lam1)*covb - 4*W1b^lam1*covb + 2*covb*lam1^2*m1bhat^2 + 4*covb*lam1*m1bhat - 4*W1b^lam1*covb*lam1*m1bhat)*s2bhat^3 + 2*covb*(covb^3*lam1^2 - 2*covb^2*lam1*m2bhat + 2*W1b^lam1*covb^2*lam1*m2bhat - 2*covb^2*lam1^2*m1bhat*m2bhat)*s2bhat)*s1bhat)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_9=sum(0) h4_10=sum(- (covb*(s2bhat^4*(- 4*lam2^2*m1bhat^2*s1bhat + 2*lam2^2*s1bhat^3) - s1bhat^3*s2bhat^2*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2)) - covb^3*(s1bhat*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2) + 2*lam2^2*s1bhat*s2bhat^2) + s1bhat^3*s2bhat^4*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + covb^2*s1bhat*s2bhat^2*(6*lam2*m1bhat + 6*lam2^2*m1bhat*m2bhat - 6*W2b^lam2*lam2*m1bhat))/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2bhat - 6*W1b^lam1*lam2 - 6*W2b^lam2*lam2 + 6*W1b^lam1*W2b^lam2*lam2 - 6*W1b^lam1*lam2^2*m2bhat)*covb^2*s1bhat*s2bhat^2 + (8*W1b^lam1*lam2^2*m1bhat - 8*lam2^2*m1bhat)*covb*s1bhat*s2bhat^4 + (2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat)*s1bhat^3*s2bhat^4) - covb*s1bhat*s2bhat^4*(4*W1b^(2*lam1)*lam2^2 - 8*W1b^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_11=sum((2*s1bhat*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_12=sum(-(2*covb*s1bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h5_1=sum(cova/(- cova^2 + s1ahat^2*s2ahat^2)) h5_2=sum((s1ahat*(2*cova^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat) - 2*cova*s2ahat^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_3=0 h5_4=0 h5_5=sum(-2/(2*s2ahat^2 - (2*cova^2)/s1ahat^2)) h5_6=sum(-(2*s1ahat^2*s2ahat*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_7=0 h5_8=0 h5_9=sum((cova^2*(lam1*m1ahat - W1a^lam1 + 1) + s1ahat^2*((lam1*m1ahat - W1a^lam1 + 1)*s2ahat^2 - 2*cova*lam1*m2ahat))/(lam1*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s1ahat^2*(2*lam1 - 2*W2a^lam2*lam1))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_10=0 h5_11=sum(-(cova*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- cova^2 + s1ahat^2*s2ahat^2))) h5_12=sum((2*s1ahat^2*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h6_1=sum((s2ahat*(2*cova^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat) - 2*cova*s1ahat^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_2=sum((lam2*((2*cova*s2ahat^3*(2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat) - 2*cova*s2ahat*(4*cova*lam1^2*m2ahat - 4*W2a^lam2*cova*lam1^2*m2ahat))*s1ahat^3 + 2*cova*s2ahat*(2*cova^2*lam1 + 2*cova^2*lam1^2*m1ahat - 2*W1a^lam1*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1^2*m1ahat + 2*W1a^lam1*W2a^lam2*cova^2*lam1)*s1ahat) - 2*cova*s1ahat^3*s2ahat*(2*cova*lam1^2 + 2*W2a^(2*lam2)*cova*lam1^2 - 4*W2a^lam2*cova*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - ((4*cova^2*lam1^2*m2ahat^2*s2ahat - 2*cova*s2ahat^3*(2*lam1*m2ahat + cova*lam1^2 + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat))*s1ahat^3 + (2*cova*(2*cova + 2*W1a^(2*lam1)*cova - 4*W1a^lam1*cova + 2*cova*lam1^2*m1ahat^2 + 4*cova*lam1*m1ahat - 4*W1a^lam1*cova*lam1*m1ahat)*s2ahat^3 + 2*cova*(cova^3*lam1^2 - 2*cova^2*lam1*m2ahat + 2*W1a^lam1*cova^2*lam1*m2ahat - 2*cova^2*lam1^2*m1ahat*m2ahat)*s2ahat)*s1ahat)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_3=0 h6_4=0 h6_5=sum(-(2*s1ahat^2*s2ahat*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_6=sum(- (3*s1ahat^6*s2ahat^2 + cova^4*(lam2^2*m1ahat^2 + lam2^2*s1ahat^2) - cova^3*(2*lam2*m1ahat*s1ahat^2 - 2*W2a^lam2*lam2*m1ahat*s1ahat^2 + 2*lam2^2*m1ahat*m2ahat*s1ahat^2) - cova*(6*lam2*m1ahat*s1ahat^4*s2ahat^2 - 6*W2a^lam2*lam2*m1ahat*s1ahat^4*s2ahat^2 + 6*lam2^2*m1ahat*m2ahat*s1ahat^4*s2ahat^2) + cova^2*(W2a^(2*lam2)*s1ahat^4 - 2*W2a^lam2*s1ahat^4 + s1ahat^4 + 2*lam2*m2ahat*s1ahat^4 + lam2^2*m2ahat^2*s1ahat^4 - 2*W2a^lam2*lam2*m2ahat*s1ahat^4 + 3*lam2^2*m1ahat^2*s1ahat^2*s2ahat^2) - 6*W2a^lam2*s1ahat^6*s2ahat^2 + 3*W2a^(2*lam2)*s1ahat^6*s2ahat^2 - lam2^2*s1ahat^6*s2ahat^4 + 3*lam2^2*m2ahat^2*s1ahat^6*s2ahat^2 + 6*lam2*m2ahat*s1ahat^6*s2ahat^2 - 6*W2a^lam2*lam2*m2ahat*s1ahat^6*s2ahat^2)/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^4*(W1a^(2*lam1)*lam2^2 - 2*W1a^lam1*lam2^2 + lam2^2) + cova^2*(3*lam2^2*s1ahat^2*s2ahat^2 + 3*W1a^(2*lam1)*lam2^2*s1ahat^2*s2ahat^2 - 6*W1a^lam1*lam2^2*s1ahat^2*s2ahat^2) - lam1*((2*W1a^lam1*lam2^2*m1ahat - 2*lam2^2*m1ahat)*cova^4 + (2*lam2*s1ahat^2 + 2*lam2^2*m2ahat*s1ahat^2 - 2*W1a^lam1*lam2*s1ahat^2 - 2*W2a^lam2*lam2*s1ahat^2 - 2*W1a^lam1*lam2^2*m2ahat*s1ahat^2 + 2*W1a^lam1*W2a^lam2*lam2*s1ahat^2)*cova^3 + (6*W1a^lam1*lam2^2*m1ahat*s1ahat^2*s2ahat^2 - 6*lam2^2*m1ahat*s1ahat^2*s2ahat^2)*cova^2 + (6*lam2*s1ahat^4*s2ahat^2 + 6*lam2^2*m2ahat*s1ahat^4*s2ahat^2 - 6*W1a^lam1*lam2*s1ahat^4*s2ahat^2 - 6*W2a^lam2*lam2*s1ahat^4*s2ahat^2 - 6*W1a^lam1*lam2^2*m2ahat*s1ahat^4*s2ahat^2 + 6*W1a^lam1*W2a^lam2*lam2*s1ahat^4*s2ahat^2)*cova))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_7=0 h6_8=0 h6_9=sum(- (cova*(s1ahat^4*(- 4*lam1^2*m2ahat^2*s2ahat + 2*lam1^2*s2ahat^3) - s1ahat^2*s2ahat^3*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2)) - cova^3*(s2ahat*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2) + 2*lam1^2*s1ahat^2*s2ahat) + s1ahat^4*s2ahat^3*(2*lam1*m2ahat + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat) + cova^2*s1ahat^2*s2ahat*(6*lam1*m2ahat + 6*lam1^2*m1ahat*m2ahat - 6*W1a^lam1*lam1*m2ahat))/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1ahat - 6*W1a^lam1*lam1 - 6*W2a^lam2*lam1 + 6*W1a^lam1*W2a^lam2*lam1 - 6*W2a^lam2*lam1^2*m1ahat)*cova^2*s1ahat^2*s2ahat + (8*W2a^lam2*lam1^2*m2ahat - 8*lam1^2*m2ahat)*cova*s1ahat^4*s2ahat + (2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat)*s1ahat^4*s2ahat^3) - cova*s1ahat^4*s2ahat*(4*W2a^(2*lam2)*lam1^2 - 8*W2a^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_10=0 h6_11=sum(-(2*cova*s2ahat*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_12=sum((2*s1ahat^2*s2ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h7_1=0 h7_2=0 h7_3=sum(covb/(- covb^2 + s1bhat^2*s2bhat^2)) h7_4=sum((s1bhat*(2*covb^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat) - 2*covb*s2bhat^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_5=0 h7_6=0 h7_7=sum(-2/(2*s2bhat^2 - (2*covb^2)/s1bhat^2)) h7_8=sum(-(2*s1bhat^2*s2bhat*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_9=0 h7_10=sum((covb^2*(lam1*m1bhat - W1b^lam1 + 1) + s1bhat^2*((lam1*m1bhat - W1b^lam1 + 1)*s2bhat^2 - 2*covb*lam1*m2bhat))/(lam1*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s1bhat^2*(2*lam1 - 2*W2b^lam2*lam1))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_11=sum(-(covb*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- covb^2 + s1bhat^2*s2bhat^2))) h7_12=sum((2*s1bhat^2*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h8_1=0 h8_2=0 h8_3=sum((s2bhat*(2*covb^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat) - 2*covb*s1bhat^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_4=sum((lam2*((2*covb*s2bhat^3*(2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat) - 2*covb*s2bhat*(4*covb*lam1^2*m2bhat - 4*W2b^lam2*covb*lam1^2*m2bhat))*s1bhat^3 + 2*covb*s2bhat*(2*covb^2*lam1 + 2*covb^2*lam1^2*m1bhat - 2*W1b^lam1*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1^2*m1bhat + 2*W1b^lam1*W2b^lam2*covb^2*lam1)*s1bhat) - 2*covb*s1bhat^3*s2bhat*(2*covb*lam1^2 + 2*W2b^(2*lam2)*covb*lam1^2 - 4*W2b^lam2*covb*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - ((4*covb^2*lam1^2*m2bhat^2*s2bhat - 2*covb*s2bhat^3*(2*lam1*m2bhat + covb*lam1^2 + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat))*s1bhat^3 + (2*covb*(2*covb + 2*W1b^(2*lam1)*covb - 4*W1b^lam1*covb + 2*covb*lam1^2*m1bhat^2 + 4*covb*lam1*m1bhat - 4*W1b^lam1*covb*lam1*m1bhat)*s2bhat^3 + 2*covb*(covb^3*lam1^2 - 2*covb^2*lam1*m2bhat + 2*W1b^lam1*covb^2*lam1*m2bhat - 2*covb^2*lam1^2*m1bhat*m2bhat)*s2bhat)*s1bhat)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_5=0 h8_6=0 h8_7=sum(-(2*s1bhat^2*s2bhat*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_8=sum(- (3*s1bhat^6*s2bhat^2 + covb^4*(lam2^2*m1bhat^2 + lam2^2*s1bhat^2) - covb^3*(2*lam2*m1bhat*s1bhat^2 - 2*W2b^lam2*lam2*m1bhat*s1bhat^2 + 2*lam2^2*m1bhat*m2bhat*s1bhat^2) - covb*(6*lam2*m1bhat*s1bhat^4*s2bhat^2 - 6*W2b^lam2*lam2*m1bhat*s1bhat^4*s2bhat^2 + 6*lam2^2*m1bhat*m2bhat*s1bhat^4*s2bhat^2) + covb^2*(W2b^(2*lam2)*s1bhat^4 - 2*W2b^lam2*s1bhat^4 + s1bhat^4 + 2*lam2*m2bhat*s1bhat^4 + lam2^2*m2bhat^2*s1bhat^4 - 2*W2b^lam2*lam2*m2bhat*s1bhat^4 + 3*lam2^2*m1bhat^2*s1bhat^2*s2bhat^2) - 6*W2b^lam2*s1bhat^6*s2bhat^2 + 3*W2b^(2*lam2)*s1bhat^6*s2bhat^2 - lam2^2*s1bhat^6*s2bhat^4 + 3*lam2^2*m2bhat^2*s1bhat^6*s2bhat^2 + 6*lam2*m2bhat*s1bhat^6*s2bhat^2 - 6*W2b^lam2*lam2*m2bhat*s1bhat^6*s2bhat^2)/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^4*(W1b^(2*lam1)*lam2^2 - 2*W1b^lam1*lam2^2 + lam2^2) + covb^2*(3*lam2^2*s1bhat^2*s2bhat^2 + 3*W1b^(2*lam1)*lam2^2*s1bhat^2*s2bhat^2 - 6*W1b^lam1*lam2^2*s1bhat^2*s2bhat^2) - lam1*((2*W1b^lam1*lam2^2*m1bhat - 2*lam2^2*m1bhat)*covb^4 + (2*lam2*s1bhat^2 + 2*lam2^2*m2bhat*s1bhat^2 - 2*W1b^lam1*lam2*s1bhat^2 - 2*W2b^lam2*lam2*s1bhat^2 - 2*W1b^lam1*lam2^2*m2bhat*s1bhat^2 + 2*W1b^lam1*W2b^lam2*lam2*s1bhat^2)*covb^3 + (6*W1b^lam1*lam2^2*m1bhat*s1bhat^2*s2bhat^2 - 6*lam2^2*m1bhat*s1bhat^2*s2bhat^2)*covb^2 + (6*lam2*s1bhat^4*s2bhat^2 + 6*lam2^2*m2bhat*s1bhat^4*s2bhat^2 - 6*W1b^lam1*lam2*s1bhat^4*s2bhat^2 - 6*W2b^lam2*lam2*s1bhat^4*s2bhat^2 - 6*W1b^lam1*lam2^2*m2bhat*s1bhat^4*s2bhat^2 + 6*W1b^lam1*W2b^lam2*lam2*s1bhat^4*s2bhat^2)*covb))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_9=0 h8_10=sum(- (covb*(s1bhat^4*(- 4*lam1^2*m2bhat^2*s2bhat + 2*lam1^2*s2bhat^3) - s1bhat^2*s2bhat^3*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2)) - covb^3*(s2bhat*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2) + 2*lam1^2*s1bhat^2*s2bhat) + s1bhat^4*s2bhat^3*(2*lam1*m2bhat + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat) + covb^2*s1bhat^2*s2bhat*(6*lam1*m2bhat + 6*lam1^2*m1bhat*m2bhat - 6*W1b^lam1*lam1*m2bhat))/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1bhat - 6*W1b^lam1*lam1 - 6*W2b^lam2*lam1 + 6*W1b^lam1*W2b^lam2*lam1 - 6*W2b^lam2*lam1^2*m1bhat)*covb^2*s1bhat^2*s2bhat + (8*W2b^lam2*lam1^2*m2bhat - 8*lam1^2*m2bhat)*covb*s1bhat^4*s2bhat + (2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat)*s1bhat^4*s2bhat^3) - covb*s1bhat^4*s2bhat*(4*W2b^(2*lam2)*lam1^2 - 8*W2b^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_11=sum(-(2*covb*s2bhat*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_12=sum((2*s1bhat^2*s2bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h9_1=sum((cova^2*(lam2*m2ahat - W2a^lam2 + 1) + s2ahat^2*((lam2*m2ahat - W2a^lam2 + 1)*s1ahat^2 - 2*cova*lam2*m1ahat))/(lam2*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s2ahat^2*(2*lam2 - 2*W1a^lam1*lam2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_2=sum(- (cova*(s2ahat^4*(- 4*lam2^2*m1ahat^2*s1ahat + 2*lam2^2*s1ahat^3) - s1ahat^3*s2ahat^2*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2)) - cova^3*(s1ahat*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2) + 2*lam2^2*s1ahat*s2ahat^2) + s1ahat^3*s2ahat^4*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + cova^2*s1ahat*s2ahat^2*(6*lam2*m1ahat + 6*lam2^2*m1ahat*m2ahat - 6*W2a^lam2*lam2*m1ahat))/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2ahat - 6*W1a^lam1*lam2 - 6*W2a^lam2*lam2 + 6*W1a^lam1*W2a^lam2*lam2 - 6*W1a^lam1*lam2^2*m2ahat)*cova^2*s1ahat*s2ahat^2 + (8*W1a^lam1*lam2^2*m1ahat - 8*lam2^2*m1ahat)*cova*s1ahat*s2ahat^4 + (2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat)*s1ahat^3*s2ahat^4) - cova*s1ahat*s2ahat^4*(4*W1a^(2*lam1)*lam2^2 - 8*W1a^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_3=0 h9_4=0 h9_5=sum((cova^2*(lam1*m1ahat - W1a^lam1 + 1) + s1ahat^2*((lam1*m1ahat - W1a^lam1 + 1)*s2ahat^2 - 2*cova*lam1*m2ahat))/(lam1*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s1ahat^2*(2*lam1 - 2*W2a^lam2*lam1))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_6=sum(- (cova*(s1ahat^4*(- 4*lam1^2*m2ahat^2*s2ahat + 2*lam1^2*s2ahat^3) - s1ahat^2*s2ahat^3*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2)) - cova^3*(s2ahat*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2) + 2*lam1^2*s1ahat^2*s2ahat) + s1ahat^4*s2ahat^3*(2*lam1*m2ahat + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat) + cova^2*s1ahat^2*s2ahat*(6*lam1*m2ahat + 6*lam1^2*m1ahat*m2ahat - 6*W1a^lam1*lam1*m2ahat))/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1ahat - 6*W1a^lam1*lam1 - 6*W2a^lam2*lam1 + 6*W1a^lam1*W2a^lam2*lam1 - 6*W2a^lam2*lam1^2*m1ahat)*cova^2*s1ahat^2*s2ahat + (8*W2a^lam2*lam1^2*m2ahat - 8*lam1^2*m2ahat)*cova*s1ahat^4*s2ahat + (2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat)*s1ahat^4*s2ahat^3) - cova*s1ahat^4*s2ahat*(4*W2a^(2*lam2)*lam1^2 - 8*W2a^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_7=0 h9_8=0 h9_9=sum(- (s1ahat^2*(cova^2*(6*lam2*m2ahat - 6*W2a^lam2 + 3*W2a^(2*lam2) + 3*lam2^2*m2ahat^2 - 6*W2a^lam2*lam2*m2ahat + 3) - cova*(6*lam2*m1ahat*s2ahat^2 - 6*W2a^lam2*lam2*m1ahat*s2ahat^2 + 6*lam2^2*m1ahat*m2ahat*s2ahat^2) + lam2^2*m1ahat^2*s2ahat^4) - cova^3*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + s1ahat^4*(W2a^(2*lam2)*s2ahat^2 - 2*W2a^lam2*s2ahat^2 + s2ahat^2 - lam2^2*s2ahat^4 + 2*lam2*m2ahat*s2ahat^2 + lam2^2*m2ahat^2*s2ahat^2 - 2*W2a^lam2*lam2*m2ahat*s2ahat^2) + cova^4*lam2^2 + 3*cova^2*lam2^2*m1ahat^2*s2ahat^2)/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^2*(3*lam2^2*s2ahat^2 - 6*W1a^lam1*lam2^2*s2ahat^2 + 3*W1a^(2*lam1)*lam2^2*s2ahat^2) + s1ahat^2*(lam2^2*s2ahat^4 - 2*W1a^lam1*lam2^2*s2ahat^4 + W1a^(2*lam1)*lam2^2*s2ahat^4) - lam1*(cova^3*(2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat) - cova^2*(6*lam2^2*m1ahat*s2ahat^2 - 6*W1a^lam1*lam2^2*m1ahat*s2ahat^2) + s1ahat^2*(cova*(6*lam2*s2ahat^2 + 6*lam2^2*m2ahat*s2ahat^2 - 6*W1a^lam1*lam2*s2ahat^2 - 6*W2a^lam2*lam2*s2ahat^2 - 6*W1a^lam1*lam2^2*m2ahat*s2ahat^2 + 6*W1a^lam1*W2a^lam2*lam2*s2ahat^2) - 2*lam2^2*m1ahat*s2ahat^4 + 2*W1a^lam1*lam2^2*m1ahat*s2ahat^4)))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_10=0 h9_11=sum(-((W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova^2*lam1 - 2*cova*lam2*s2ahat^2 + lam1*s1ahat^2*s2ahat^2 - W2a^lam2*cova^2*lam1 + 2*W1a^lam1*cova*lam2*s2ahat^2 + cova^2*lam1*lam2*m2ahat - W2a^lam2*lam1*s1ahat^2*s2ahat^2 + lam1*lam2*m2ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_12=sum(-((W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova^2*lam2 - 2*cova*lam1*s1ahat^2 + lam2*s1ahat^2*s2ahat^2 - W1a^lam1*cova^2*lam2 + 2*W2a^lam2*cova*lam1*s1ahat^2 + cova^2*lam1*lam2*m1ahat - W1a^lam1*lam2*s1ahat^2*s2ahat^2 + lam1*lam2*m1ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h10_1=0 h10_2=0 h10_3=sum((covb^2*(lam2*m2bhat - W2b^lam2 + 1) + s2bhat^2*((lam2*m2bhat - W2b^lam2 + 1)*s1bhat^2 - 2*covb*lam2*m1bhat))/(lam2*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s2bhat^2*(2*lam2 - 2*W1b^lam1*lam2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_4=sum(- (covb*(s2bhat^4*(- 4*lam2^2*m1bhat^2*s1bhat + 2*lam2^2*s1bhat^3) - s1bhat^3*s2bhat^2*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2)) - covb^3*(s1bhat*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2) + 2*lam2^2*s1bhat*s2bhat^2) + s1bhat^3*s2bhat^4*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + covb^2*s1bhat*s2bhat^2*(6*lam2*m1bhat + 6*lam2^2*m1bhat*m2bhat - 6*W2b^lam2*lam2*m1bhat))/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2bhat - 6*W1b^lam1*lam2 - 6*W2b^lam2*lam2 + 6*W1b^lam1*W2b^lam2*lam2 - 6*W1b^lam1*lam2^2*m2bhat)*covb^2*s1bhat*s2bhat^2 + (8*W1b^lam1*lam2^2*m1bhat - 8*lam2^2*m1bhat)*covb*s1bhat*s2bhat^4 + (2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat)*s1bhat^3*s2bhat^4) - covb*s1bhat*s2bhat^4*(4*W1b^(2*lam1)*lam2^2 - 8*W1b^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_5=0 h10_6=0 h10_7=sum((covb^2*(lam1*m1bhat - W1b^lam1 + 1) + s1bhat^2*((lam1*m1bhat - W1b^lam1 + 1)*s2bhat^2 - 2*covb*lam1*m2bhat))/(lam1*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s1bhat^2*(2*lam1 - 2*W2b^lam2*lam1))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_8=sum(- (covb*(s1bhat^4*(- 4*lam1^2*m2bhat^2*s2bhat + 2*lam1^2*s2bhat^3) - s1bhat^2*s2bhat^3*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2)) - covb^3*(s2bhat*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2) + 2*lam1^2*s1bhat^2*s2bhat) + s1bhat^4*s2bhat^3*(2*lam1*m2bhat + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat) + covb^2*s1bhat^2*s2bhat*(6*lam1*m2bhat + 6*lam1^2*m1bhat*m2bhat - 6*W1b^lam1*lam1*m2bhat))/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1bhat - 6*W1b^lam1*lam1 - 6*W2b^lam2*lam1 + 6*W1b^lam1*W2b^lam2*lam1 - 6*W2b^lam2*lam1^2*m1bhat)*covb^2*s1bhat^2*s2bhat + (8*W2b^lam2*lam1^2*m2bhat - 8*lam1^2*m2bhat)*covb*s1bhat^4*s2bhat + (2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat)*s1bhat^4*s2bhat^3) - covb*s1bhat^4*s2bhat*(4*W2b^(2*lam2)*lam1^2 - 8*W2b^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_9=0 h10_10=sum(- (s1bhat^2*(covb^2*(6*lam2*m2bhat - 6*W2b^lam2 + 3*W2b^(2*lam2) + 3*lam2^2*m2bhat^2 - 6*W2b^lam2*lam2*m2bhat + 3) - covb*(6*lam2*m1bhat*s2bhat^2 - 6*W2b^lam2*lam2*m1bhat*s2bhat^2 + 6*lam2^2*m1bhat*m2bhat*s2bhat^2) + lam2^2*m1bhat^2*s2bhat^4) - covb^3*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + s1bhat^4*(W2b^(2*lam2)*s2bhat^2 - 2*W2b^lam2*s2bhat^2 + s2bhat^2 - lam2^2*s2bhat^4 + 2*lam2*m2bhat*s2bhat^2 + lam2^2*m2bhat^2*s2bhat^2 - 2*W2b^lam2*lam2*m2bhat*s2bhat^2) + covb^4*lam2^2 + 3*covb^2*lam2^2*m1bhat^2*s2bhat^2)/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^2*(3*lam2^2*s2bhat^2 - 6*W1b^lam1*lam2^2*s2bhat^2 + 3*W1b^(2*lam1)*lam2^2*s2bhat^2) + s1bhat^2*(lam2^2*s2bhat^4 - 2*W1b^lam1*lam2^2*s2bhat^4 + W1b^(2*lam1)*lam2^2*s2bhat^4) - lam1*(covb^3*(2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat) - covb^2*(6*lam2^2*m1bhat*s2bhat^2 - 6*W1b^lam1*lam2^2*m1bhat*s2bhat^2) + s1bhat^2*(covb*(6*lam2*s2bhat^2 + 6*lam2^2*m2bhat*s2bhat^2 - 6*W1b^lam1*lam2*s2bhat^2 - 6*W2b^lam2*lam2*s2bhat^2 - 6*W1b^lam1*lam2^2*m2bhat*s2bhat^2 + 6*W1b^lam1*W2b^lam2*lam2*s2bhat^2) - 2*lam2^2*m1bhat*s2bhat^4 + 2*W1b^lam1*lam2^2*m1bhat*s2bhat^4)))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_11=sum(-((W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb^2*lam1 - 2*covb*lam2*s2bhat^2 + lam1*s1bhat^2*s2bhat^2 - W2b^lam2*covb^2*lam1 + 2*W1b^lam1*covb*lam2*s2bhat^2 + covb^2*lam1*lam2*m2bhat - W2b^lam2*lam1*s1bhat^2*s2bhat^2 + lam1*lam2*m2bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_12=sum(-((W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb^2*lam2 - 2*covb*lam1*s1bhat^2 + lam2*s1bhat^2*s2bhat^2 - W1b^lam1*covb^2*lam2 + 2*W2b^lam2*covb*lam1*s1bhat^2 + covb^2*lam1*lam2*m1bhat - W1b^lam1*lam2*s1bhat^2*s2bhat^2 + lam1*lam2*m1bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_1=sum((2*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h11_2=sum((2*s1ahat*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_3=sum((2*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h11_4=sum((2*s1bhat*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_5=sum(-(cova*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- cova^2 + s1ahat^2*s2ahat^2))) h11_6=sum(-(2*cova*s2ahat*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_7=sum(-(covb*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- covb^2 + s1bhat^2*s2bhat^2))) h11_8=sum(-(2*covb*s2bhat*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_9=sum(-((W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova^2*lam1 - 2*cova*lam2*s2ahat^2 + lam1*s1ahat^2*s2ahat^2 - W2a^lam2*cova^2*lam1 + 2*W1a^lam1*cova*lam2*s2ahat^2 + cova^2*lam1*lam2*m2ahat - W2a^lam2*lam1*s1ahat^2*s2ahat^2 + lam1*lam2*m2ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_10=sum(-((W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb^2*lam1 - 2*covb*lam2*s2bhat^2 + lam1*s1bhat^2*s2bhat^2 - W2b^lam2*covb^2*lam1 + 2*W1b^lam1*covb*lam2*s2bhat^2 + covb^2*lam1*lam2*m2bhat - W2b^lam2*lam1*s1bhat^2*s2bhat^2 + lam1*lam2*m2bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_11=sum(((2*((W1a^lam1-1)/lam1^2-(W1a^lam1*log(W1a))/lam1)^2)/s1ahat^2-(2*(m1ahat-(W1a^lam1-1)/lam1)*((2*W1a^lam1-2)/lam1^3-(2*W1a^lam1*log(W1a))/lam1^2+(W1a^lam1*log(W1a)^2)/lam1))/s1ahat^2+(2*cova*(m2ahat-(W2a^lam2-1)/lam2)*((2*W1a^lam1-2)/lam1^3-(2*W1a^lam1*log(W1a))/lam1^2+(W1a^lam1*log(W1a)^2)/lam1))/(s1ahat^2*s2ahat^2))/((2*cova^2)/(s1ahat^2*s2ahat^2)-2))+sum(((2*((W1b^lam1-1)/lam1^2-(W1b^lam1*log(W1b))/lam1)^2)/s1bhat^2-(2*(m1bhat-(W1b^lam1-1)/lam1)*((2*W1b^lam1-2)/lam1^3-(2*W1b^lam1*log(W1b))/lam1^2+(W1b^lam1*log(W1b)^2)/lam1))/s1bhat^2+(2*covb*(m2bhat-(W2b^lam2-1)/lam2)*((2*W1b^lam1-2)/lam1^3-(2*W1b^lam1*log(W1b))/lam1^2+(W1b^lam1*log(W1b)^2)/lam1))/(s1bhat^2*s2bhat^2))/((2*covb^2)/(s1bhat^2*s2bhat^2)-2)) h11_12=sum((2*cova*(W1a^lam1*lam1*log(W1a)-W1a^lam1+1)*(W2a^lam2*lam2*log(W2a)-W2a^lam2+1))/(lam1^2*lam2^2*(-2*cova^2+2*s1ahat^2*s2ahat^2)))+sum((2*covb*(W1b^lam1*lam1*log(W1b)-W1b^lam1+1)*(W2b^lam2*lam2*log(W2b)-W2b^lam2+1))/(lam1^2*lam2^2*(-2*covb^2+2*s1bhat^2*s2bhat^2))) h12_1=sum(-(cova*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- cova^2 + s1ahat^2*s2ahat^2))) h12_2=sum(-(2*cova*s1ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_3=sum(-(covb*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- covb^2 + s1bhat^2*s2bhat^2))) h12_4=sum(-(2*covb*s1bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_5=sum((2*s1ahat^2*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h12_6=sum((2*s1ahat^2*s2ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_7=sum((2*s1bhat^2*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h12_8=sum((2*s1bhat^2*s2bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_9=sum(-((W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova^2*lam2 - 2*cova*lam1*s1ahat^2 + lam2*s1ahat^2*s2ahat^2 - W1a^lam1*cova^2*lam2 + 2*W2a^lam2*cova*lam1*s1ahat^2 + cova^2*lam1*lam2*m1ahat - W1a^lam1*lam2*s1ahat^2*s2ahat^2 + lam1*lam2*m1ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_10=sum(-((W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb^2*lam2 - 2*covb*lam1*s1bhat^2 + lam2*s1bhat^2*s2bhat^2 - W1b^lam1*covb^2*lam2 + 2*W2b^lam2*covb*lam1*s1bhat^2 + covb^2*lam1*lam2*m1bhat - W1b^lam1*lam2*s1bhat^2*s2bhat^2 + lam1*lam2*m1bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_11=sum((2*cova*(W1a^lam1*lam1*log(W1a)-W1a^lam1+1)*(W2a^lam2*lam2*log(W2a)-W2a^lam2+1))/(lam1^2*lam2^2*(-2*cova^2+2*s1ahat^2*s2ahat^2)))+sum((2*covb*(W1b^lam1*lam1*log(W1b)-W1b^lam1+1)*(W2b^lam2*lam2*log(W2b)-W2b^lam2+1))/(lam1^2*lam2^2*(-2*covb^2+2*s1bhat^2*s2bhat^2))) h12_12=sum(((2*((W2a^lam2-1)/lam2^2-(W2a^lam2*log(W2a))/lam2)^2)/s2ahat^2-(2*(m2ahat-(W2a^lam2-1)/lam2)*((2*W2a^lam2-2)/lam2^3-(2*W2a^lam2*log(W2a))/lam2^2+(W2a^lam2*log(W2a)^2)/lam2))/s2ahat^2+(2*cova*(m1ahat-(W1a^lam1-1)/lam1)*((2*W2a^lam2-2)/lam2^3-(2*W2a^lam2*log(W2a))/lam2^2+(W2a^lam2*log(W2a)^2)/lam2))/(s1ahat^2*s2ahat^2))/((2*cova^2)/(s1ahat^2*s2ahat^2)-2))+sum(((2*((W2b^lam2-1)/lam2^2-(W2b^lam2*log(W2b))/lam2)^2)/s2bhat^2-(2*(m2bhat-(W2b^lam2-1)/lam2)*((2*W2b^lam2-2)/lam2^3-(2*W2b^lam2*log(W2b))/lam2^2+(W2b^lam2*log(W2b)^2)/lam2))/s2bhat^2+(2*covb*(m1bhat-(W1b^lam1-1)/lam1)*((2*W2b^lam2-2)/lam2^3-(2*W2b^lam2*log(W2b))/lam2^2+(W2b^lam2*log(W2b)^2)/lam2))/(s1bhat^2*s2bhat^2))/((2*covb^2)/(s1bhat^2*s2bhat^2)-2)) h1r=c(h1_1,h1_2,h1_3,h1_4,h1_5,h1_6,h1_7,h1_8,h1_9,h1_10,h1_11,h1_12) h2r=c(h2_1,h2_2,h2_3,h2_4,h2_5,h2_6,h2_7,h2_8,h2_9,h2_10,h2_11,h2_12) h3r=c(h3_1,h3_2,h3_3,h3_4,h3_5,h3_6,h3_7,h3_8,h3_9,h3_10,h3_11,h3_12) h4r=c(h4_1,h4_2,h4_3,h4_4,h4_5,h4_6,h4_7,h4_8,h4_9,h4_10,h4_11,h4_12) h5r=c(h5_1,h5_2,h5_3,h5_4,h5_5,h5_6,h5_7,h5_8,h5_9,h5_10,h5_11,h5_12) h6r=c(h6_1,h6_2,h6_3,h6_4,h6_5,h6_6,h6_7,h6_8,h6_9,h6_10,h6_11,h6_12) h7r=c(h7_1,h7_2,h7_3,h7_4,h7_5,h7_6,h7_7,h7_8,h7_9,h7_10,h7_11,h7_12) h8r=c(h8_1,h8_2,h8_3,h8_4,h8_5,h8_6,h8_7,h8_8,h8_9,h8_10,h8_11,h8_12) h9r=c(h9_1,h9_2,h9_3,h9_4,h9_5,h9_6,h9_7,h9_8,h9_9,h9_10,h9_11,h9_12) h10r=c(h10_1,h10_2,h10_3,h10_4,h10_5,h10_6,h10_7,h10_8,h10_9,h10_10,h10_11,h10_12) h11r=c(h11_1,h11_2,h11_3,h11_4,h11_5,h11_6,h11_7,h11_8,h11_9,h11_10,h11_11,h11_12) h12r=c(h12_1,h12_2,h12_3,h12_4,h12_5,h12_6,h12_7,h12_8,h12_9,h12_10,h12_11,h12_12) HCF=rbind(h1r,h2r,h3r,h4r,h5r,h6r,h7r,h8r,h9r,h10r,h11r,h12r) HCF[1,2]=0;HCF[1,6]=0;HCF[1,9]=0; HCF[2,1]=0;HCF[2,5]=0; HCF[3,4]=0;HCF[3,8]=0;HCF[3,10]=0;HCF[4,3]=0; HCF[4,7]=0; HCF[5,2]=0;HCF[5,6]=0;HCF[5,9]=0; HCF[6,1]=0;HCF[6,5]=0; HCF[7,4]=0;HCF[7,8]=0;HCF[7,10]=0; HCF[8,3]=0;HCF[8,7]=0; HCF[9,1]=0;HCF[9,5]=0; HCF[10,3]=0; HCF[10,7]=0; HCF[1,1]=na*HCF[1,1]; HCF[1,5]=na*HCF[1,5]; HCF[5,1]=na*HCF[5,1]; HCF[3,3]=nb*HCF[3,3]; HCF[3,7]=nb*HCF[3,7]; HCF[7,3]=nb*HCF[7,3]; HCF[7,7]=nb*HCF[7,7]; HCF[5,5]=na*HCF[5,5]; VV=inv(-HCF) ############################################################################# #DEALING WITH ORIGINAL AUCS ############################################################################# AUC1 = erfc((2^(1/2)*(m1ahat - m1bhat))/(2*(s1ahat^2 + s1bhat^2)^(1/2)))/2 dm1a = -(2^(1/2)*exp(-(m1ahat - m1bhat)^2/(2*(s1ahat^2 + s1bhat^2))))/(2*pi^(1/2)*(s1ahat^2 + s1bhat^2)^(1/2)) dm1b = (2^(1/2)*exp(-(m1ahat - m1bhat)^2/(2*(s1ahat^2 + s1bhat^2))))/(2*pi^(1/2)*(s1ahat^2 + s1bhat^2)^(1/2)) ds1a = (2^(1/2)*s1ahat*exp(-(m1ahat - m1bhat)^2/(2*(s1ahat^2 + s1bhat^2)))*(m1ahat - m1bhat))/(2*pi^(1/2)*(s1ahat^2 + s1bhat^2)^(3/2)) ds1b = (2^(1/2)*s1bhat*exp(-(m1ahat - m1bhat)^2/(2*(s1ahat^2 + s1bhat^2)))*(m1ahat - m1bhat))/(2*pi^(1/2)*(s1ahat^2 + s1bhat^2)^(3/2)) AUC2 = erfc((2^(1/2)*(m2ahat - m2bhat))/(2*(s2ahat^2 + s2bhat^2)^(1/2)))/2 dm2a = -(2^(1/2)*exp(-(m2ahat - m2bhat)^2/(2*(s2ahat^2 + s2bhat^2))))/(2*pi^(1/2)*(s2ahat^2 + s2bhat^2)^(1/2)) dm2b = (2^(1/2)*exp(-(m2ahat - m2bhat)^2/(2*(s2ahat^2 + s2bhat^2))))/(2*pi^(1/2)*(s2ahat^2 + s2bhat^2)^(1/2)) ds2a = (2^(1/2)*s2ahat*exp(-(m2ahat - m2bhat)^2/(2*(s2ahat^2 + s2bhat^2)))*(m2ahat - m2bhat))/(2*pi^(1/2)*(s2ahat^2 + s2bhat^2)^(3/2)) ds2b = (2^(1/2)*s2bhat*exp(-(m2ahat - m2bhat)^2/(2*(s2ahat^2 + s2bhat^2)))*(m2ahat - m2bhat))/(2*pi^(1/2)*(s2ahat^2 + s2bhat^2)^(3/2)) varAUC1=c(dm1a, ds1a, dm1b, ds1b)%*%VV[1:4,1:4]%*%t(t(c(dm1a, ds1a, dm1b, ds1b))); varAUC2=c(dm2a, ds2a, dm2b, ds2b)%*%VV[5:8,5:8]%*%t(t(c(dm2a, ds2a, dm2b, ds2b))); varAUC1=c(varAUC1) varAUC2=c(varAUC2) V=zeros(8,8); covAUC=c(dm1a, ds1a, dm1b, ds1b, 0, 0, 0, 0)%*%VV[1:8,1:8]%*%t(t(c(0, 0, 0, 0, dm2a, ds2a, dm2b, ds2b))) covAUC=c(covAUC) ZdAUC=(AUC2-AUC1)/(sqrt(varAUC1+varAUC2-2*covAUC)) SEdAUC=sqrt(varAUC1+varAUC2-2*covAUC) CIdiffdAUC=c(AUC2-AUC1-pmalpha*SEdAUC, AUC2-AUC1+pmalpha*SEdAUC) pval2tdAUC=2*pnorm(-abs(ZdAUC)) ############################################################################# #PROBIT VERSIONS ############################################################################# #norminv versions AUC1= ((m1bhat-m1ahat)/s1bhat)/(sqrt(1+(s1ahat/s1bhat)^2)); AUC2= ((m2bhat-m2ahat)/s2bhat)/(sqrt(1+(s2ahat/s2bhat)^2)); AUC1MC=AUC1 AUC2MC=AUC2 AUC1original=pnorm(AUC1) AUC2original=pnorm(AUC2) #====================norminv AUC1============================== dm1a =-1/(s1bhat*sqrt(1+(s1ahat/s1bhat)^2)); ds1a = (s1ahat*(m1ahat-m1bhat))/(s1bhat^3*(1+(s1ahat/s1bhat)^2)^(3/2)) dm1b =1/(s1bhat*sqrt(1+(s1ahat/s1bhat)^2)); ds1b =((m1ahat-m1bhat))/(s1bhat^2*(1+(s1ahat/s1bhat)^2)^(3/2)) varAUC1=c(dm1a, ds1a, dm1b, ds1b)%*%VV[1:4,1:4]%*%t(t(c(dm1a, ds1a, dm1b, ds1b))) #====================norminv AUC2============================== dm2a =-1/(s2bhat*sqrt(1+(s2ahat/s2bhat)^2)); ds2a = (s2ahat*(m2ahat-m2bhat))/(s2bhat^3*(1+(s2ahat/s2bhat)^2)^(3/2)) dm2b =1/(s2bhat*sqrt(1+(s2ahat/s2bhat)^2)); ds2b =((m2ahat-m2bhat))/(s2bhat^2*(1+(s2ahat/s2bhat)^2)^(3/2)) varAUC2=c(dm2a, ds2a, dm2b, ds2b)%*%VV[5:8,5:8]%*%t(t(c(dm2a, ds2a, dm2b, ds2b))); #====================norminv cov(AUC1,AUC2)============================== V=zeros(8,8); covAUC=c(dm1a, ds1a, dm1b, ds1b, 0, 0, 0, 0)%*%VV[1:8,1:8]%*%t(t(c(0, 0, 0, 0, dm2a, ds2a, dm2b, ds2b))) covAUC=c(covAUC) Z=(AUC2MC-AUC1MC)/(sqrt(varAUC1+varAUC2-2*covAUC)) SE=sqrt(varAUC1+varAUC2-2*covAUC) CIdiff=c(AUC2MC-AUC1MC-pmalpha*SE, AUC2MC-AUC1MC+pmalpha*SE) pval2t=2*pnorm(-abs(Z)) #################### IF PLOTS ARE REQUESTED ################################# #====TWO ROC FUNCTIONS: ONE IS THE BOXCOX AND THE OTHER THE REFERENCE LINE================ roc1<-function(t){ na = length(W1alam) nb = length(W1blam) s1alam=sqrt( var(W1alam)*(na-1)/na ) s1blam=sqrt( var(W1blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W1alam), sd=s1alam), mean=mean(W1blam), sd=s1blam) } roc2<-function(t){ na = length(W2alam) nb = length(W2blam) s2alam=sqrt( var(W2alam)*(na-1)/na ) s2blam=sqrt( var(W2blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W2alam), sd=s2alam), mean=mean(W2blam), sd=s2blam) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } #================IF PLOTS ARE REQUESTED PLOT THE ROCS== if (plots=="on") { # x11() plot(linspace(0,1,1000),roc1(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="red") lines(linspace(0,1,1000),roc2(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="black") lines(linspace(0,1,1000),linspace(0,1,1000),type="l", lty=2) legend("bottomright", legend=c(paste("ROC for Marker 1 with AUC =", formattable(AUC1original, digits = 4, format = "f")), paste("ROC for Marker 2 with AUC =", formattable(AUC2original, digits = 4, format = "f")), paste("P-value for the probit AUC difference:", formattable(pval2t, digits = 4, format = "f"), " ")), col=c("red", "black", "white"), lty=c(1, 1, NA), pch = c(NA, NA, NA), cex=0.8) } ############################################################################### ############################################################################### #================OUTPUT ARGUMENTS====================== res <- data.frame(AUC1 = formattable(AUC1original, digits = 4, format = "f"), AUC2 = formattable(AUC2original, digits = 4, format = "f"), diff = formattable(AUC2original - AUC1original, digits = 4, format = "f"), p_value_probit = formattable(pval2t, digits = 4, format = "f"), p_value = formattable(pval2tdAUC, digits = 4, format = "f"), ci_ll = formattable(CIdiffdAUC[1], digits = 4, format = "f"), ci_ul = formattable(CIdiffdAUC[2], digits = 4, format = "f")) rownames(res) <- c("Estimates:") colnames(res) <- c("AUC 1", "AUC 2", "Diff", "P-Val (Probit)", "P-Val", "CI (LL)", "CI (UL)") res <- formattable(as.matrix(res), digits = 4, format = "f") res #return(list(AUCmarker1=AUC1original,AUCmarker2=AUC2original, pvalue_difference= pval2t, CI_difference= pnorm(CIdiff), rocbc1=roc1, rocbc2=roc2)) return(list(resultstable=res, AUCmarker1=AUC1original, AUCmarker2=AUC2original, pvalue_probit_difference= pval2t, pvalue_difference= pval2tdAUC, CI_difference= CIdiffdAUC, roc1=roc1, roc2=roc2, transx1=W1alam, transy1=W1blam, transformation.parameter.1 = lam[1], transx2=W2alam, transy2=W2blam, transformation.parameter.2 = lam[2])) } } ``` ```{r, echo=FALSE, eval=TRUE} comparebcJ <-function (marker1, marker2, D, alpha, plots){ if ((length(marker1) != length(D)) | (length(marker2) != length(D))) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (alpha <= 0 | alpha >= 1) { stop("ERROR: The level of significance, alpha, should be set between 0 and 1. A common choice is 0.05.") } else if (sum(is.na(marker1)) > 0 | sum(is.na(marker2)) > 0 | sum(is.na(D)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if ((sum(marker1 < 0) > 0) | (sum(marker2 < 2)) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { erf <- function (x) 2 * pnorm(x * sqrt(2)) - 1 erfc <- function (x) 2 * pnorm(x * sqrt(2), lower.tail = FALSE) erfinv <- function (x) qnorm((1 + x)/2)/sqrt(2) erfcinv <- function (x) qnorm(x/2, lower.tail = FALSE)/sqrt(2) pmalpha=qnorm(1-alpha/2); if (plots!="on"){plots="off"} W1a=marker1[D==0] W1b=marker1[D==1] W2a=marker2[D==0] W2b=marker2[D==1] Wa=cbind(W1a,W2a); Wb=cbind(W1b,W2b); na=length(W1a) nb=length(W1b) likbox2D<-function(W1a,W1b,W2a,W2b,lam){ na=length(W1a); nb=length(W1b); out=c(); #for (i in 1:length(h)){ W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) #cova=1/na*sum((W1alam-mean(W1alam))*(W2alam-mean(W2alam))) #covb=1/nb*sum((W1blam-mean(W1blam))*(W2blam-mean(W2blam))) cova= cov(Walam)*(na-1)/na; covb= cov(Wblam)*(nb-1)/nb; mualam=c(mean(W1alam), mean(W2alam)) mublam=c(mean(W1blam), mean(W2blam)) #dmvn(c(0,0), mu, mcov, log = F) # out= (-sum(log(dmvn(Walam,mualam,cova)))-sum(log(dmvn(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b))); out= -sum(log(dmvnorm(Walam,mualam,cova)))-sum(log(dmvnorm(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b)); #out1=-sum(log(dmvnorm(Walam,mualam,cova))) #out2=-sum(log(dmvnorm(Wblam,mublam,covb))) #out3=-(lam[1]-1)*sum(log(W1a)) #out4=-(lam[2]-1)*sum(log(W2a)) #out5=-(lam[1]-1)*sum(log(W1b)) #out6=-(lam[2]-1)*sum(log(W2b)) #out=c(out,out1,out2,out3,out4,out5,out6, cova, covb) #out= (-sum(log(mvnpdf2([W1alam(g(1)) W2alam(g(2))],[mean(W1alam(g(1))) mean(W2alam(g(2)))],[std(W1alam(g(1)),1).^2 cova;cova (std(W2alam(g(2)),1)).^2])))+... # -sum(log(mvnpdf2([W1blam(g(1)) W2blam(g(2))],[mean(W1blam(g(1))) mean(W2blam(g(2)))],[std(W1blam(g(1)),1).^2 covb;covb (std(W2blam(g(2)),1)).^2])))+... # -(g(1)-1).*sum(log(W1a))-(g(2)-1).*sum(log(W2a))-(g(1)-1).*sum(log(W1b))-(g(2)-1).*sum(log(W2b))); return(out) } lam=c(2,2) likbox2D(W1a,W1b,W2a,W2b,lam) logL<-function(h){ likbox2D(W1a,W1b,W2a,W2b,h) } # init=c(0.8,0.8) # lam=fminsearch(logL,init) # lam=c(lam$optbase$xopt) lam<- optim(c(1,1),logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) lam #init=c(-10,10) #lam=optimize(logL,init) #lam #f <- function (x, a) (x - a)^2 #xmin <- optimize(f, c(0, 1), tol = 0.0001, a = 1/3) #xmin out <- optim(c(1,1), logL, method = "Nelder-Mead") lam = out$par ################################### W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) m1ahat=mean(W1alam) m1bhat=mean(W1blam) m2ahat=mean(W2alam) m2bhat=mean(W2blam) s1ahat=sqrt( var(W1alam)*(na-1)/na ) s1bhat=sqrt( var(W1blam)*(nb-1)/nb ) s2ahat=sqrt( var(W2alam)*(na-1)/na ) s2bhat=sqrt( var(W2blam)*(nb-1)/nb ) cova= cov(Walam)*(na-1)/na;cova=cova[1,2]; covb= cov(Wblam)*(nb-1)/nb;covb=covb[1,2]; lam1=lam[1] lam2=lam[2] ################################## h1_1=sum(-2/(2*s1ahat^2 - (2*cova^2)/s2ahat^2)) h1_2=sum(-(2*s1ahat*s2ahat^2*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_3=0 h1_4=0; h1_5=sum(cova/(- cova^2 + s1ahat^2*s2ahat^2)); h1_6=sum((s2ahat*(2*cova^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat) - 2*cova*s1ahat^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_7=0 h1_8=0 h1_9=sum((cova^2*(lam2*m2ahat - W2a^lam2 + 1) + s2ahat^2*((lam2*m2ahat - W2a^lam2 + 1)*s1ahat^2 - 2*cova*lam2*m1ahat))/(lam2*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s2ahat^2*(2*lam2 - 2*W1a^lam1*lam2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_10=0 h1_11=sum((2*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h1_12=sum(-(cova*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- cova^2 + s1ahat^2*s2ahat^2))) h2_1=sum(-(2*s1ahat*s2ahat^2*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_2=sum(- (3*s1ahat^2*s2ahat^6 + cova^4*(lam1^2*m2ahat^2 + lam1^2*s2ahat^2) - cova^3*(2*lam1*m2ahat*s2ahat^2 - 2*W1a^lam1*lam1*m2ahat*s2ahat^2 + 2*lam1^2*m1ahat*m2ahat*s2ahat^2) - cova*(6*lam1*m2ahat*s1ahat^2*s2ahat^4 - 6*W1a^lam1*lam1*m2ahat*s1ahat^2*s2ahat^4 + 6*lam1^2*m1ahat*m2ahat*s1ahat^2*s2ahat^4) + cova^2*(W1a^(2*lam1)*s2ahat^4 - 2*W1a^lam1*s2ahat^4 + s2ahat^4 + 2*lam1*m1ahat*s2ahat^4 + lam1^2*m1ahat^2*s2ahat^4 - 2*W1a^lam1*lam1*m1ahat*s2ahat^4 + 3*lam1^2*m2ahat^2*s1ahat^2*s2ahat^2) - 6*W1a^lam1*s1ahat^2*s2ahat^6 + 3*W1a^(2*lam1)*s1ahat^2*s2ahat^6 - lam1^2*s1ahat^4*s2ahat^6 + 3*lam1^2*m1ahat^2*s1ahat^2*s2ahat^6 + 6*lam1*m1ahat*s1ahat^2*s2ahat^6 - 6*W1a^lam1*lam1*m1ahat*s1ahat^2*s2ahat^6)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^4*(W2a^(2*lam2)*lam1^2 - 2*W2a^lam2*lam1^2 + lam1^2) + cova^2*(3*lam1^2*s1ahat^2*s2ahat^2 + 3*W2a^(2*lam2)*lam1^2*s1ahat^2*s2ahat^2 - 6*W2a^lam2*lam1^2*s1ahat^2*s2ahat^2) - lam2*((2*W2a^lam2*lam1^2*m2ahat - 2*lam1^2*m2ahat)*cova^4 + (2*lam1*s2ahat^2 + 2*lam1^2*m1ahat*s2ahat^2 - 2*W1a^lam1*lam1*s2ahat^2 - 2*W2a^lam2*lam1*s2ahat^2 - 2*W2a^lam2*lam1^2*m1ahat*s2ahat^2 + 2*W1a^lam1*W2a^lam2*lam1*s2ahat^2)*cova^3 + (6*W2a^lam2*lam1^2*m2ahat*s1ahat^2*s2ahat^2 - 6*lam1^2*m2ahat*s1ahat^2*s2ahat^2)*cova^2 + (6*lam1*s1ahat^2*s2ahat^4 + 6*lam1^2*m1ahat*s1ahat^2*s2ahat^4 - 6*W1a^lam1*lam1*s1ahat^2*s2ahat^4 - 6*W2a^lam2*lam1*s1ahat^2*s2ahat^4 - 6*W2a^lam2*lam1^2*m1ahat*s1ahat^2*s2ahat^4 + 6*W1a^lam1*W2a^lam2*lam1*s1ahat^2*s2ahat^4)*cova))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_3=sum(0) h2_4=sum(0) h2_5=sum((s1ahat*(2*cova^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat) - 2*cova*s2ahat^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_6=sum((lam2*((2*cova*s2ahat^3*(2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat) - 2*cova*s2ahat*(4*cova*lam1^2*m2ahat - 4*W2a^lam2*cova*lam1^2*m2ahat))*s1ahat^3 + 2*cova*s2ahat*(2*cova^2*lam1 + 2*cova^2*lam1^2*m1ahat - 2*W1a^lam1*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1^2*m1ahat + 2*W1a^lam1*W2a^lam2*cova^2*lam1)*s1ahat) - 2*cova*s1ahat^3*s2ahat*(2*cova*lam1^2 + 2*W2a^(2*lam2)*cova*lam1^2 - 4*W2a^lam2*cova*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - ((4*cova^2*lam1^2*m2ahat^2*s2ahat - 2*cova*s2ahat^3*(2*lam1*m2ahat + cova*lam1^2 + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat))*s1ahat^3 + (2*cova*(2*cova + 2*W1a^(2*lam1)*cova - 4*W1a^lam1*cova + 2*cova*lam1^2*m1ahat^2 + 4*cova*lam1*m1ahat - 4*W1a^lam1*cova*lam1*m1ahat)*s2ahat^3 + 2*cova*(cova^3*lam1^2 - 2*cova^2*lam1*m2ahat + 2*W1a^lam1*cova^2*lam1*m2ahat - 2*cova^2*lam1^2*m1ahat*m2ahat)*s2ahat)*s1ahat)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_7=sum(0) h2_8=sum(0) h2_9=sum(- (cova*(s2ahat^4*(- 4*lam2^2*m1ahat^2*s1ahat + 2*lam2^2*s1ahat^3) - s1ahat^3*s2ahat^2*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2)) - cova^3*(s1ahat*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2) + 2*lam2^2*s1ahat*s2ahat^2) + s1ahat^3*s2ahat^4*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + cova^2*s1ahat*s2ahat^2*(6*lam2*m1ahat + 6*lam2^2*m1ahat*m2ahat - 6*W2a^lam2*lam2*m1ahat))/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2ahat - 6*W1a^lam1*lam2 - 6*W2a^lam2*lam2 + 6*W1a^lam1*W2a^lam2*lam2 - 6*W1a^lam1*lam2^2*m2ahat)*cova^2*s1ahat*s2ahat^2 + (8*W1a^lam1*lam2^2*m1ahat - 8*lam2^2*m1ahat)*cova*s1ahat*s2ahat^4 + (2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat)*s1ahat^3*s2ahat^4) - cova*s1ahat*s2ahat^4*(4*W1a^(2*lam1)*lam2^2 - 8*W1a^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_10=sum(0) h2_11=sum((2*s1ahat*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_12=sum(-(2*cova*s1ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h3_1=sum(0) h3_2=sum(0) h3_3=sum(-2/(2*s1bhat^2 - (2*covb^2)/s2bhat^2)) h3_4=sum(-(2*s1bhat*s2bhat^2*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_5=sum(0) h3_6=sum(0) h3_7=sum(covb/(- covb^2 + s1bhat^2*s2bhat^2)) h3_8=sum((s2bhat*(2*covb^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat) - 2*covb*s1bhat^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_9=sum(0) h3_10=sum((covb^2*(lam2*m2bhat - W2b^lam2 + 1) + s2bhat^2*((lam2*m2bhat - W2b^lam2 + 1)*s1bhat^2 - 2*covb*lam2*m1bhat))/(lam2*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s2bhat^2*(2*lam2 - 2*W1b^lam1*lam2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_11=sum((2*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h3_12=sum(-(covb*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- covb^2 + s1bhat^2*s2bhat^2))) h4_1=sum(0) h4_2=sum(0) h4_3=sum(-(2*s1bhat*s2bhat^2*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_4=sum(- (3*s1bhat^2*s2bhat^6 + covb^4*(lam1^2*m2bhat^2 + lam1^2*s2bhat^2) - covb^3*(2*lam1*m2bhat*s2bhat^2 - 2*W1b^lam1*lam1*m2bhat*s2bhat^2 + 2*lam1^2*m1bhat*m2bhat*s2bhat^2) - covb*(6*lam1*m2bhat*s1bhat^2*s2bhat^4 - 6*W1b^lam1*lam1*m2bhat*s1bhat^2*s2bhat^4 + 6*lam1^2*m1bhat*m2bhat*s1bhat^2*s2bhat^4) + covb^2*(W1b^(2*lam1)*s2bhat^4 - 2*W1b^lam1*s2bhat^4 + s2bhat^4 + 2*lam1*m1bhat*s2bhat^4 + lam1^2*m1bhat^2*s2bhat^4 - 2*W1b^lam1*lam1*m1bhat*s2bhat^4 + 3*lam1^2*m2bhat^2*s1bhat^2*s2bhat^2) - 6*W1b^lam1*s1bhat^2*s2bhat^6 + 3*W1b^(2*lam1)*s1bhat^2*s2bhat^6 - lam1^2*s1bhat^4*s2bhat^6 + 3*lam1^2*m1bhat^2*s1bhat^2*s2bhat^6 + 6*lam1*m1bhat*s1bhat^2*s2bhat^6 - 6*W1b^lam1*lam1*m1bhat*s1bhat^2*s2bhat^6)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^4*(W2b^(2*lam2)*lam1^2 - 2*W2b^lam2*lam1^2 + lam1^2) + covb^2*(3*lam1^2*s1bhat^2*s2bhat^2 + 3*W2b^(2*lam2)*lam1^2*s1bhat^2*s2bhat^2 - 6*W2b^lam2*lam1^2*s1bhat^2*s2bhat^2) - lam2*((2*W2b^lam2*lam1^2*m2bhat - 2*lam1^2*m2bhat)*covb^4 + (2*lam1*s2bhat^2 + 2*lam1^2*m1bhat*s2bhat^2 - 2*W1b^lam1*lam1*s2bhat^2 - 2*W2b^lam2*lam1*s2bhat^2 - 2*W2b^lam2*lam1^2*m1bhat*s2bhat^2 + 2*W1b^lam1*W2b^lam2*lam1*s2bhat^2)*covb^3 + (6*W2b^lam2*lam1^2*m2bhat*s1bhat^2*s2bhat^2 - 6*lam1^2*m2bhat*s1bhat^2*s2bhat^2)*covb^2 + (6*lam1*s1bhat^2*s2bhat^4 + 6*lam1^2*m1bhat*s1bhat^2*s2bhat^4 - 6*W1b^lam1*lam1*s1bhat^2*s2bhat^4 - 6*W2b^lam2*lam1*s1bhat^2*s2bhat^4 - 6*W2b^lam2*lam1^2*m1bhat*s1bhat^2*s2bhat^4 + 6*W1b^lam1*W2b^lam2*lam1*s1bhat^2*s2bhat^4)*covb))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_5=sum(0) h4_6=sum(0) h4_7=sum((s1bhat*(2*covb^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat) - 2*covb*s2bhat^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_8=sum((lam2*((2*covb*s2bhat^3*(2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat) - 2*covb*s2bhat*(4*covb*lam1^2*m2bhat - 4*W2b^lam2*covb*lam1^2*m2bhat))*s1bhat^3 + 2*covb*s2bhat*(2*covb^2*lam1 + 2*covb^2*lam1^2*m1bhat - 2*W1b^lam1*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1^2*m1bhat + 2*W1b^lam1*W2b^lam2*covb^2*lam1)*s1bhat) - 2*covb*s1bhat^3*s2bhat*(2*covb*lam1^2 + 2*W2b^(2*lam2)*covb*lam1^2 - 4*W2b^lam2*covb*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - ((4*covb^2*lam1^2*m2bhat^2*s2bhat - 2*covb*s2bhat^3*(2*lam1*m2bhat + covb*lam1^2 + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat))*s1bhat^3 + (2*covb*(2*covb + 2*W1b^(2*lam1)*covb - 4*W1b^lam1*covb + 2*covb*lam1^2*m1bhat^2 + 4*covb*lam1*m1bhat - 4*W1b^lam1*covb*lam1*m1bhat)*s2bhat^3 + 2*covb*(covb^3*lam1^2 - 2*covb^2*lam1*m2bhat + 2*W1b^lam1*covb^2*lam1*m2bhat - 2*covb^2*lam1^2*m1bhat*m2bhat)*s2bhat)*s1bhat)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_9=sum(0) h4_10=sum(- (covb*(s2bhat^4*(- 4*lam2^2*m1bhat^2*s1bhat + 2*lam2^2*s1bhat^3) - s1bhat^3*s2bhat^2*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2)) - covb^3*(s1bhat*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2) + 2*lam2^2*s1bhat*s2bhat^2) + s1bhat^3*s2bhat^4*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + covb^2*s1bhat*s2bhat^2*(6*lam2*m1bhat + 6*lam2^2*m1bhat*m2bhat - 6*W2b^lam2*lam2*m1bhat))/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2bhat - 6*W1b^lam1*lam2 - 6*W2b^lam2*lam2 + 6*W1b^lam1*W2b^lam2*lam2 - 6*W1b^lam1*lam2^2*m2bhat)*covb^2*s1bhat*s2bhat^2 + (8*W1b^lam1*lam2^2*m1bhat - 8*lam2^2*m1bhat)*covb*s1bhat*s2bhat^4 + (2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat)*s1bhat^3*s2bhat^4) - covb*s1bhat*s2bhat^4*(4*W1b^(2*lam1)*lam2^2 - 8*W1b^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_11=sum((2*s1bhat*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_12=sum(-(2*covb*s1bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h5_1=sum(cova/(- cova^2 + s1ahat^2*s2ahat^2)) h5_2=sum((s1ahat*(2*cova^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat) - 2*cova*s2ahat^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_3=0 h5_4=0 h5_5=sum(-2/(2*s2ahat^2 - (2*cova^2)/s1ahat^2)) h5_6=sum(-(2*s1ahat^2*s2ahat*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_7=0 h5_8=0 h5_9=sum((cova^2*(lam1*m1ahat - W1a^lam1 + 1) + s1ahat^2*((lam1*m1ahat - W1a^lam1 + 1)*s2ahat^2 - 2*cova*lam1*m2ahat))/(lam1*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s1ahat^2*(2*lam1 - 2*W2a^lam2*lam1))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_10=0 h5_11=sum(-(cova*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- cova^2 + s1ahat^2*s2ahat^2))) h5_12=sum((2*s1ahat^2*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h6_1=sum((s2ahat*(2*cova^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat) - 2*cova*s1ahat^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_2=sum((lam2*((2*cova*s2ahat^3*(2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat) - 2*cova*s2ahat*(4*cova*lam1^2*m2ahat - 4*W2a^lam2*cova*lam1^2*m2ahat))*s1ahat^3 + 2*cova*s2ahat*(2*cova^2*lam1 + 2*cova^2*lam1^2*m1ahat - 2*W1a^lam1*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1^2*m1ahat + 2*W1a^lam1*W2a^lam2*cova^2*lam1)*s1ahat) - 2*cova*s1ahat^3*s2ahat*(2*cova*lam1^2 + 2*W2a^(2*lam2)*cova*lam1^2 - 4*W2a^lam2*cova*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - ((4*cova^2*lam1^2*m2ahat^2*s2ahat - 2*cova*s2ahat^3*(2*lam1*m2ahat + cova*lam1^2 + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat))*s1ahat^3 + (2*cova*(2*cova + 2*W1a^(2*lam1)*cova - 4*W1a^lam1*cova + 2*cova*lam1^2*m1ahat^2 + 4*cova*lam1*m1ahat - 4*W1a^lam1*cova*lam1*m1ahat)*s2ahat^3 + 2*cova*(cova^3*lam1^2 - 2*cova^2*lam1*m2ahat + 2*W1a^lam1*cova^2*lam1*m2ahat - 2*cova^2*lam1^2*m1ahat*m2ahat)*s2ahat)*s1ahat)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_3=0 h6_4=0 h6_5=sum(-(2*s1ahat^2*s2ahat*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_6=sum(- (3*s1ahat^6*s2ahat^2 + cova^4*(lam2^2*m1ahat^2 + lam2^2*s1ahat^2) - cova^3*(2*lam2*m1ahat*s1ahat^2 - 2*W2a^lam2*lam2*m1ahat*s1ahat^2 + 2*lam2^2*m1ahat*m2ahat*s1ahat^2) - cova*(6*lam2*m1ahat*s1ahat^4*s2ahat^2 - 6*W2a^lam2*lam2*m1ahat*s1ahat^4*s2ahat^2 + 6*lam2^2*m1ahat*m2ahat*s1ahat^4*s2ahat^2) + cova^2*(W2a^(2*lam2)*s1ahat^4 - 2*W2a^lam2*s1ahat^4 + s1ahat^4 + 2*lam2*m2ahat*s1ahat^4 + lam2^2*m2ahat^2*s1ahat^4 - 2*W2a^lam2*lam2*m2ahat*s1ahat^4 + 3*lam2^2*m1ahat^2*s1ahat^2*s2ahat^2) - 6*W2a^lam2*s1ahat^6*s2ahat^2 + 3*W2a^(2*lam2)*s1ahat^6*s2ahat^2 - lam2^2*s1ahat^6*s2ahat^4 + 3*lam2^2*m2ahat^2*s1ahat^6*s2ahat^2 + 6*lam2*m2ahat*s1ahat^6*s2ahat^2 - 6*W2a^lam2*lam2*m2ahat*s1ahat^6*s2ahat^2)/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^4*(W1a^(2*lam1)*lam2^2 - 2*W1a^lam1*lam2^2 + lam2^2) + cova^2*(3*lam2^2*s1ahat^2*s2ahat^2 + 3*W1a^(2*lam1)*lam2^2*s1ahat^2*s2ahat^2 - 6*W1a^lam1*lam2^2*s1ahat^2*s2ahat^2) - lam1*((2*W1a^lam1*lam2^2*m1ahat - 2*lam2^2*m1ahat)*cova^4 + (2*lam2*s1ahat^2 + 2*lam2^2*m2ahat*s1ahat^2 - 2*W1a^lam1*lam2*s1ahat^2 - 2*W2a^lam2*lam2*s1ahat^2 - 2*W1a^lam1*lam2^2*m2ahat*s1ahat^2 + 2*W1a^lam1*W2a^lam2*lam2*s1ahat^2)*cova^3 + (6*W1a^lam1*lam2^2*m1ahat*s1ahat^2*s2ahat^2 - 6*lam2^2*m1ahat*s1ahat^2*s2ahat^2)*cova^2 + (6*lam2*s1ahat^4*s2ahat^2 + 6*lam2^2*m2ahat*s1ahat^4*s2ahat^2 - 6*W1a^lam1*lam2*s1ahat^4*s2ahat^2 - 6*W2a^lam2*lam2*s1ahat^4*s2ahat^2 - 6*W1a^lam1*lam2^2*m2ahat*s1ahat^4*s2ahat^2 + 6*W1a^lam1*W2a^lam2*lam2*s1ahat^4*s2ahat^2)*cova))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_7=0 h6_8=0 h6_9=sum(- (cova*(s1ahat^4*(- 4*lam1^2*m2ahat^2*s2ahat + 2*lam1^2*s2ahat^3) - s1ahat^2*s2ahat^3*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2)) - cova^3*(s2ahat*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2) + 2*lam1^2*s1ahat^2*s2ahat) + s1ahat^4*s2ahat^3*(2*lam1*m2ahat + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat) + cova^2*s1ahat^2*s2ahat*(6*lam1*m2ahat + 6*lam1^2*m1ahat*m2ahat - 6*W1a^lam1*lam1*m2ahat))/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1ahat - 6*W1a^lam1*lam1 - 6*W2a^lam2*lam1 + 6*W1a^lam1*W2a^lam2*lam1 - 6*W2a^lam2*lam1^2*m1ahat)*cova^2*s1ahat^2*s2ahat + (8*W2a^lam2*lam1^2*m2ahat - 8*lam1^2*m2ahat)*cova*s1ahat^4*s2ahat + (2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat)*s1ahat^4*s2ahat^3) - cova*s1ahat^4*s2ahat*(4*W2a^(2*lam2)*lam1^2 - 8*W2a^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_10=0 h6_11=sum(-(2*cova*s2ahat*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_12=sum((2*s1ahat^2*s2ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h7_1=0 h7_2=0 h7_3=sum(covb/(- covb^2 + s1bhat^2*s2bhat^2)) h7_4=sum((s1bhat*(2*covb^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat) - 2*covb*s2bhat^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_5=0 h7_6=0 h7_7=sum(-2/(2*s2bhat^2 - (2*covb^2)/s1bhat^2)) h7_8=sum(-(2*s1bhat^2*s2bhat*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_9=0 h7_10=sum((covb^2*(lam1*m1bhat - W1b^lam1 + 1) + s1bhat^2*((lam1*m1bhat - W1b^lam1 + 1)*s2bhat^2 - 2*covb*lam1*m2bhat))/(lam1*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s1bhat^2*(2*lam1 - 2*W2b^lam2*lam1))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_11=sum(-(covb*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- covb^2 + s1bhat^2*s2bhat^2))) h7_12=sum((2*s1bhat^2*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h8_1=0 h8_2=0 h8_3=sum((s2bhat*(2*covb^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat) - 2*covb*s1bhat^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_4=sum((lam2*((2*covb*s2bhat^3*(2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat) - 2*covb*s2bhat*(4*covb*lam1^2*m2bhat - 4*W2b^lam2*covb*lam1^2*m2bhat))*s1bhat^3 + 2*covb*s2bhat*(2*covb^2*lam1 + 2*covb^2*lam1^2*m1bhat - 2*W1b^lam1*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1^2*m1bhat + 2*W1b^lam1*W2b^lam2*covb^2*lam1)*s1bhat) - 2*covb*s1bhat^3*s2bhat*(2*covb*lam1^2 + 2*W2b^(2*lam2)*covb*lam1^2 - 4*W2b^lam2*covb*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - ((4*covb^2*lam1^2*m2bhat^2*s2bhat - 2*covb*s2bhat^3*(2*lam1*m2bhat + covb*lam1^2 + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat))*s1bhat^3 + (2*covb*(2*covb + 2*W1b^(2*lam1)*covb - 4*W1b^lam1*covb + 2*covb*lam1^2*m1bhat^2 + 4*covb*lam1*m1bhat - 4*W1b^lam1*covb*lam1*m1bhat)*s2bhat^3 + 2*covb*(covb^3*lam1^2 - 2*covb^2*lam1*m2bhat + 2*W1b^lam1*covb^2*lam1*m2bhat - 2*covb^2*lam1^2*m1bhat*m2bhat)*s2bhat)*s1bhat)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_5=0 h8_6=0 h8_7=sum(-(2*s1bhat^2*s2bhat*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_8=sum(- (3*s1bhat^6*s2bhat^2 + covb^4*(lam2^2*m1bhat^2 + lam2^2*s1bhat^2) - covb^3*(2*lam2*m1bhat*s1bhat^2 - 2*W2b^lam2*lam2*m1bhat*s1bhat^2 + 2*lam2^2*m1bhat*m2bhat*s1bhat^2) - covb*(6*lam2*m1bhat*s1bhat^4*s2bhat^2 - 6*W2b^lam2*lam2*m1bhat*s1bhat^4*s2bhat^2 + 6*lam2^2*m1bhat*m2bhat*s1bhat^4*s2bhat^2) + covb^2*(W2b^(2*lam2)*s1bhat^4 - 2*W2b^lam2*s1bhat^4 + s1bhat^4 + 2*lam2*m2bhat*s1bhat^4 + lam2^2*m2bhat^2*s1bhat^4 - 2*W2b^lam2*lam2*m2bhat*s1bhat^4 + 3*lam2^2*m1bhat^2*s1bhat^2*s2bhat^2) - 6*W2b^lam2*s1bhat^6*s2bhat^2 + 3*W2b^(2*lam2)*s1bhat^6*s2bhat^2 - lam2^2*s1bhat^6*s2bhat^4 + 3*lam2^2*m2bhat^2*s1bhat^6*s2bhat^2 + 6*lam2*m2bhat*s1bhat^6*s2bhat^2 - 6*W2b^lam2*lam2*m2bhat*s1bhat^6*s2bhat^2)/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^4*(W1b^(2*lam1)*lam2^2 - 2*W1b^lam1*lam2^2 + lam2^2) + covb^2*(3*lam2^2*s1bhat^2*s2bhat^2 + 3*W1b^(2*lam1)*lam2^2*s1bhat^2*s2bhat^2 - 6*W1b^lam1*lam2^2*s1bhat^2*s2bhat^2) - lam1*((2*W1b^lam1*lam2^2*m1bhat - 2*lam2^2*m1bhat)*covb^4 + (2*lam2*s1bhat^2 + 2*lam2^2*m2bhat*s1bhat^2 - 2*W1b^lam1*lam2*s1bhat^2 - 2*W2b^lam2*lam2*s1bhat^2 - 2*W1b^lam1*lam2^2*m2bhat*s1bhat^2 + 2*W1b^lam1*W2b^lam2*lam2*s1bhat^2)*covb^3 + (6*W1b^lam1*lam2^2*m1bhat*s1bhat^2*s2bhat^2 - 6*lam2^2*m1bhat*s1bhat^2*s2bhat^2)*covb^2 + (6*lam2*s1bhat^4*s2bhat^2 + 6*lam2^2*m2bhat*s1bhat^4*s2bhat^2 - 6*W1b^lam1*lam2*s1bhat^4*s2bhat^2 - 6*W2b^lam2*lam2*s1bhat^4*s2bhat^2 - 6*W1b^lam1*lam2^2*m2bhat*s1bhat^4*s2bhat^2 + 6*W1b^lam1*W2b^lam2*lam2*s1bhat^4*s2bhat^2)*covb))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_9=0 h8_10=sum(- (covb*(s1bhat^4*(- 4*lam1^2*m2bhat^2*s2bhat + 2*lam1^2*s2bhat^3) - s1bhat^2*s2bhat^3*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2)) - covb^3*(s2bhat*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2) + 2*lam1^2*s1bhat^2*s2bhat) + s1bhat^4*s2bhat^3*(2*lam1*m2bhat + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat) + covb^2*s1bhat^2*s2bhat*(6*lam1*m2bhat + 6*lam1^2*m1bhat*m2bhat - 6*W1b^lam1*lam1*m2bhat))/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1bhat - 6*W1b^lam1*lam1 - 6*W2b^lam2*lam1 + 6*W1b^lam1*W2b^lam2*lam1 - 6*W2b^lam2*lam1^2*m1bhat)*covb^2*s1bhat^2*s2bhat + (8*W2b^lam2*lam1^2*m2bhat - 8*lam1^2*m2bhat)*covb*s1bhat^4*s2bhat + (2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat)*s1bhat^4*s2bhat^3) - covb*s1bhat^4*s2bhat*(4*W2b^(2*lam2)*lam1^2 - 8*W2b^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_11=sum(-(2*covb*s2bhat*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_12=sum((2*s1bhat^2*s2bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h9_1=sum((cova^2*(lam2*m2ahat - W2a^lam2 + 1) + s2ahat^2*((lam2*m2ahat - W2a^lam2 + 1)*s1ahat^2 - 2*cova*lam2*m1ahat))/(lam2*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s2ahat^2*(2*lam2 - 2*W1a^lam1*lam2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_2=sum(- (cova*(s2ahat^4*(- 4*lam2^2*m1ahat^2*s1ahat + 2*lam2^2*s1ahat^3) - s1ahat^3*s2ahat^2*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2)) - cova^3*(s1ahat*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2) + 2*lam2^2*s1ahat*s2ahat^2) + s1ahat^3*s2ahat^4*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + cova^2*s1ahat*s2ahat^2*(6*lam2*m1ahat + 6*lam2^2*m1ahat*m2ahat - 6*W2a^lam2*lam2*m1ahat))/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2ahat - 6*W1a^lam1*lam2 - 6*W2a^lam2*lam2 + 6*W1a^lam1*W2a^lam2*lam2 - 6*W1a^lam1*lam2^2*m2ahat)*cova^2*s1ahat*s2ahat^2 + (8*W1a^lam1*lam2^2*m1ahat - 8*lam2^2*m1ahat)*cova*s1ahat*s2ahat^4 + (2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat)*s1ahat^3*s2ahat^4) - cova*s1ahat*s2ahat^4*(4*W1a^(2*lam1)*lam2^2 - 8*W1a^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_3=0 h9_4=0 h9_5=sum((cova^2*(lam1*m1ahat - W1a^lam1 + 1) + s1ahat^2*((lam1*m1ahat - W1a^lam1 + 1)*s2ahat^2 - 2*cova*lam1*m2ahat))/(lam1*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s1ahat^2*(2*lam1 - 2*W2a^lam2*lam1))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_6=sum(- (cova*(s1ahat^4*(- 4*lam1^2*m2ahat^2*s2ahat + 2*lam1^2*s2ahat^3) - s1ahat^2*s2ahat^3*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2)) - cova^3*(s2ahat*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2) + 2*lam1^2*s1ahat^2*s2ahat) + s1ahat^4*s2ahat^3*(2*lam1*m2ahat + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat) + cova^2*s1ahat^2*s2ahat*(6*lam1*m2ahat + 6*lam1^2*m1ahat*m2ahat - 6*W1a^lam1*lam1*m2ahat))/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1ahat - 6*W1a^lam1*lam1 - 6*W2a^lam2*lam1 + 6*W1a^lam1*W2a^lam2*lam1 - 6*W2a^lam2*lam1^2*m1ahat)*cova^2*s1ahat^2*s2ahat + (8*W2a^lam2*lam1^2*m2ahat - 8*lam1^2*m2ahat)*cova*s1ahat^4*s2ahat + (2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat)*s1ahat^4*s2ahat^3) - cova*s1ahat^4*s2ahat*(4*W2a^(2*lam2)*lam1^2 - 8*W2a^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_7=0 h9_8=0 h9_9=sum(- (s1ahat^2*(cova^2*(6*lam2*m2ahat - 6*W2a^lam2 + 3*W2a^(2*lam2) + 3*lam2^2*m2ahat^2 - 6*W2a^lam2*lam2*m2ahat + 3) - cova*(6*lam2*m1ahat*s2ahat^2 - 6*W2a^lam2*lam2*m1ahat*s2ahat^2 + 6*lam2^2*m1ahat*m2ahat*s2ahat^2) + lam2^2*m1ahat^2*s2ahat^4) - cova^3*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + s1ahat^4*(W2a^(2*lam2)*s2ahat^2 - 2*W2a^lam2*s2ahat^2 + s2ahat^2 - lam2^2*s2ahat^4 + 2*lam2*m2ahat*s2ahat^2 + lam2^2*m2ahat^2*s2ahat^2 - 2*W2a^lam2*lam2*m2ahat*s2ahat^2) + cova^4*lam2^2 + 3*cova^2*lam2^2*m1ahat^2*s2ahat^2)/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^2*(3*lam2^2*s2ahat^2 - 6*W1a^lam1*lam2^2*s2ahat^2 + 3*W1a^(2*lam1)*lam2^2*s2ahat^2) + s1ahat^2*(lam2^2*s2ahat^4 - 2*W1a^lam1*lam2^2*s2ahat^4 + W1a^(2*lam1)*lam2^2*s2ahat^4) - lam1*(cova^3*(2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat) - cova^2*(6*lam2^2*m1ahat*s2ahat^2 - 6*W1a^lam1*lam2^2*m1ahat*s2ahat^2) + s1ahat^2*(cova*(6*lam2*s2ahat^2 + 6*lam2^2*m2ahat*s2ahat^2 - 6*W1a^lam1*lam2*s2ahat^2 - 6*W2a^lam2*lam2*s2ahat^2 - 6*W1a^lam1*lam2^2*m2ahat*s2ahat^2 + 6*W1a^lam1*W2a^lam2*lam2*s2ahat^2) - 2*lam2^2*m1ahat*s2ahat^4 + 2*W1a^lam1*lam2^2*m1ahat*s2ahat^4)))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_10=0 h9_11=sum(-((W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova^2*lam1 - 2*cova*lam2*s2ahat^2 + lam1*s1ahat^2*s2ahat^2 - W2a^lam2*cova^2*lam1 + 2*W1a^lam1*cova*lam2*s2ahat^2 + cova^2*lam1*lam2*m2ahat - W2a^lam2*lam1*s1ahat^2*s2ahat^2 + lam1*lam2*m2ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_12=sum(-((W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova^2*lam2 - 2*cova*lam1*s1ahat^2 + lam2*s1ahat^2*s2ahat^2 - W1a^lam1*cova^2*lam2 + 2*W2a^lam2*cova*lam1*s1ahat^2 + cova^2*lam1*lam2*m1ahat - W1a^lam1*lam2*s1ahat^2*s2ahat^2 + lam1*lam2*m1ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h10_1=0 h10_2=0 h10_3=sum((covb^2*(lam2*m2bhat - W2b^lam2 + 1) + s2bhat^2*((lam2*m2bhat - W2b^lam2 + 1)*s1bhat^2 - 2*covb*lam2*m1bhat))/(lam2*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s2bhat^2*(2*lam2 - 2*W1b^lam1*lam2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_4=sum(- (covb*(s2bhat^4*(- 4*lam2^2*m1bhat^2*s1bhat + 2*lam2^2*s1bhat^3) - s1bhat^3*s2bhat^2*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2)) - covb^3*(s1bhat*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2) + 2*lam2^2*s1bhat*s2bhat^2) + s1bhat^3*s2bhat^4*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + covb^2*s1bhat*s2bhat^2*(6*lam2*m1bhat + 6*lam2^2*m1bhat*m2bhat - 6*W2b^lam2*lam2*m1bhat))/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2bhat - 6*W1b^lam1*lam2 - 6*W2b^lam2*lam2 + 6*W1b^lam1*W2b^lam2*lam2 - 6*W1b^lam1*lam2^2*m2bhat)*covb^2*s1bhat*s2bhat^2 + (8*W1b^lam1*lam2^2*m1bhat - 8*lam2^2*m1bhat)*covb*s1bhat*s2bhat^4 + (2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat)*s1bhat^3*s2bhat^4) - covb*s1bhat*s2bhat^4*(4*W1b^(2*lam1)*lam2^2 - 8*W1b^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_5=0 h10_6=0 h10_7=sum((covb^2*(lam1*m1bhat - W1b^lam1 + 1) + s1bhat^2*((lam1*m1bhat - W1b^lam1 + 1)*s2bhat^2 - 2*covb*lam1*m2bhat))/(lam1*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s1bhat^2*(2*lam1 - 2*W2b^lam2*lam1))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_8=sum(- (covb*(s1bhat^4*(- 4*lam1^2*m2bhat^2*s2bhat + 2*lam1^2*s2bhat^3) - s1bhat^2*s2bhat^3*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2)) - covb^3*(s2bhat*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2) + 2*lam1^2*s1bhat^2*s2bhat) + s1bhat^4*s2bhat^3*(2*lam1*m2bhat + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat) + covb^2*s1bhat^2*s2bhat*(6*lam1*m2bhat + 6*lam1^2*m1bhat*m2bhat - 6*W1b^lam1*lam1*m2bhat))/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1bhat - 6*W1b^lam1*lam1 - 6*W2b^lam2*lam1 + 6*W1b^lam1*W2b^lam2*lam1 - 6*W2b^lam2*lam1^2*m1bhat)*covb^2*s1bhat^2*s2bhat + (8*W2b^lam2*lam1^2*m2bhat - 8*lam1^2*m2bhat)*covb*s1bhat^4*s2bhat + (2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat)*s1bhat^4*s2bhat^3) - covb*s1bhat^4*s2bhat*(4*W2b^(2*lam2)*lam1^2 - 8*W2b^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_9=0 h10_10=sum(- (s1bhat^2*(covb^2*(6*lam2*m2bhat - 6*W2b^lam2 + 3*W2b^(2*lam2) + 3*lam2^2*m2bhat^2 - 6*W2b^lam2*lam2*m2bhat + 3) - covb*(6*lam2*m1bhat*s2bhat^2 - 6*W2b^lam2*lam2*m1bhat*s2bhat^2 + 6*lam2^2*m1bhat*m2bhat*s2bhat^2) + lam2^2*m1bhat^2*s2bhat^4) - covb^3*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + s1bhat^4*(W2b^(2*lam2)*s2bhat^2 - 2*W2b^lam2*s2bhat^2 + s2bhat^2 - lam2^2*s2bhat^4 + 2*lam2*m2bhat*s2bhat^2 + lam2^2*m2bhat^2*s2bhat^2 - 2*W2b^lam2*lam2*m2bhat*s2bhat^2) + covb^4*lam2^2 + 3*covb^2*lam2^2*m1bhat^2*s2bhat^2)/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^2*(3*lam2^2*s2bhat^2 - 6*W1b^lam1*lam2^2*s2bhat^2 + 3*W1b^(2*lam1)*lam2^2*s2bhat^2) + s1bhat^2*(lam2^2*s2bhat^4 - 2*W1b^lam1*lam2^2*s2bhat^4 + W1b^(2*lam1)*lam2^2*s2bhat^4) - lam1*(covb^3*(2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat) - covb^2*(6*lam2^2*m1bhat*s2bhat^2 - 6*W1b^lam1*lam2^2*m1bhat*s2bhat^2) + s1bhat^2*(covb*(6*lam2*s2bhat^2 + 6*lam2^2*m2bhat*s2bhat^2 - 6*W1b^lam1*lam2*s2bhat^2 - 6*W2b^lam2*lam2*s2bhat^2 - 6*W1b^lam1*lam2^2*m2bhat*s2bhat^2 + 6*W1b^lam1*W2b^lam2*lam2*s2bhat^2) - 2*lam2^2*m1bhat*s2bhat^4 + 2*W1b^lam1*lam2^2*m1bhat*s2bhat^4)))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_11=sum(-((W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb^2*lam1 - 2*covb*lam2*s2bhat^2 + lam1*s1bhat^2*s2bhat^2 - W2b^lam2*covb^2*lam1 + 2*W1b^lam1*covb*lam2*s2bhat^2 + covb^2*lam1*lam2*m2bhat - W2b^lam2*lam1*s1bhat^2*s2bhat^2 + lam1*lam2*m2bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_12=sum(-((W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb^2*lam2 - 2*covb*lam1*s1bhat^2 + lam2*s1bhat^2*s2bhat^2 - W1b^lam1*covb^2*lam2 + 2*W2b^lam2*covb*lam1*s1bhat^2 + covb^2*lam1*lam2*m1bhat - W1b^lam1*lam2*s1bhat^2*s2bhat^2 + lam1*lam2*m1bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_1=sum((2*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h11_2=sum((2*s1ahat*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_3=sum((2*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h11_4=sum((2*s1bhat*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_5=sum(-(cova*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- cova^2 + s1ahat^2*s2ahat^2))) h11_6=sum(-(2*cova*s2ahat*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_7=sum(-(covb*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- covb^2 + s1bhat^2*s2bhat^2))) h11_8=sum(-(2*covb*s2bhat*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_9=sum(-((W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova^2*lam1 - 2*cova*lam2*s2ahat^2 + lam1*s1ahat^2*s2ahat^2 - W2a^lam2*cova^2*lam1 + 2*W1a^lam1*cova*lam2*s2ahat^2 + cova^2*lam1*lam2*m2ahat - W2a^lam2*lam1*s1ahat^2*s2ahat^2 + lam1*lam2*m2ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_10=sum(-((W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb^2*lam1 - 2*covb*lam2*s2bhat^2 + lam1*s1bhat^2*s2bhat^2 - W2b^lam2*covb^2*lam1 + 2*W1b^lam1*covb*lam2*s2bhat^2 + covb^2*lam1*lam2*m2bhat - W2b^lam2*lam1*s1bhat^2*s2bhat^2 + lam1*lam2*m2bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_11=sum(((2*((W1a^lam1-1)/lam1^2-(W1a^lam1*log(W1a))/lam1)^2)/s1ahat^2-(2*(m1ahat-(W1a^lam1-1)/lam1)*((2*W1a^lam1-2)/lam1^3-(2*W1a^lam1*log(W1a))/lam1^2+(W1a^lam1*log(W1a)^2)/lam1))/s1ahat^2+(2*cova*(m2ahat-(W2a^lam2-1)/lam2)*((2*W1a^lam1-2)/lam1^3-(2*W1a^lam1*log(W1a))/lam1^2+(W1a^lam1*log(W1a)^2)/lam1))/(s1ahat^2*s2ahat^2))/((2*cova^2)/(s1ahat^2*s2ahat^2)-2))+sum(((2*((W1b^lam1-1)/lam1^2-(W1b^lam1*log(W1b))/lam1)^2)/s1bhat^2-(2*(m1bhat-(W1b^lam1-1)/lam1)*((2*W1b^lam1-2)/lam1^3-(2*W1b^lam1*log(W1b))/lam1^2+(W1b^lam1*log(W1b)^2)/lam1))/s1bhat^2+(2*covb*(m2bhat-(W2b^lam2-1)/lam2)*((2*W1b^lam1-2)/lam1^3-(2*W1b^lam1*log(W1b))/lam1^2+(W1b^lam1*log(W1b)^2)/lam1))/(s1bhat^2*s2bhat^2))/((2*covb^2)/(s1bhat^2*s2bhat^2)-2)) h11_12=sum((2*cova*(W1a^lam1*lam1*log(W1a)-W1a^lam1+1)*(W2a^lam2*lam2*log(W2a)-W2a^lam2+1))/(lam1^2*lam2^2*(-2*cova^2+2*s1ahat^2*s2ahat^2)))+sum((2*covb*(W1b^lam1*lam1*log(W1b)-W1b^lam1+1)*(W2b^lam2*lam2*log(W2b)-W2b^lam2+1))/(lam1^2*lam2^2*(-2*covb^2+2*s1bhat^2*s2bhat^2))) h12_1=sum(-(cova*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- cova^2 + s1ahat^2*s2ahat^2))) h12_2=sum(-(2*cova*s1ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_3=sum(-(covb*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- covb^2 + s1bhat^2*s2bhat^2))) h12_4=sum(-(2*covb*s1bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_5=sum((2*s1ahat^2*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h12_6=sum((2*s1ahat^2*s2ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_7=sum((2*s1bhat^2*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h12_8=sum((2*s1bhat^2*s2bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_9=sum(-((W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova^2*lam2 - 2*cova*lam1*s1ahat^2 + lam2*s1ahat^2*s2ahat^2 - W1a^lam1*cova^2*lam2 + 2*W2a^lam2*cova*lam1*s1ahat^2 + cova^2*lam1*lam2*m1ahat - W1a^lam1*lam2*s1ahat^2*s2ahat^2 + lam1*lam2*m1ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_10=sum(-((W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb^2*lam2 - 2*covb*lam1*s1bhat^2 + lam2*s1bhat^2*s2bhat^2 - W1b^lam1*covb^2*lam2 + 2*W2b^lam2*covb*lam1*s1bhat^2 + covb^2*lam1*lam2*m1bhat - W1b^lam1*lam2*s1bhat^2*s2bhat^2 + lam1*lam2*m1bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_11=sum((2*cova*(W1a^lam1*lam1*log(W1a)-W1a^lam1+1)*(W2a^lam2*lam2*log(W2a)-W2a^lam2+1))/(lam1^2*lam2^2*(-2*cova^2+2*s1ahat^2*s2ahat^2)))+sum((2*covb*(W1b^lam1*lam1*log(W1b)-W1b^lam1+1)*(W2b^lam2*lam2*log(W2b)-W2b^lam2+1))/(lam1^2*lam2^2*(-2*covb^2+2*s1bhat^2*s2bhat^2))) h12_12=sum(((2*((W2a^lam2-1)/lam2^2-(W2a^lam2*log(W2a))/lam2)^2)/s2ahat^2-(2*(m2ahat-(W2a^lam2-1)/lam2)*((2*W2a^lam2-2)/lam2^3-(2*W2a^lam2*log(W2a))/lam2^2+(W2a^lam2*log(W2a)^2)/lam2))/s2ahat^2+(2*cova*(m1ahat-(W1a^lam1-1)/lam1)*((2*W2a^lam2-2)/lam2^3-(2*W2a^lam2*log(W2a))/lam2^2+(W2a^lam2*log(W2a)^2)/lam2))/(s1ahat^2*s2ahat^2))/((2*cova^2)/(s1ahat^2*s2ahat^2)-2))+sum(((2*((W2b^lam2-1)/lam2^2-(W2b^lam2*log(W2b))/lam2)^2)/s2bhat^2-(2*(m2bhat-(W2b^lam2-1)/lam2)*((2*W2b^lam2-2)/lam2^3-(2*W2b^lam2*log(W2b))/lam2^2+(W2b^lam2*log(W2b)^2)/lam2))/s2bhat^2+(2*covb*(m1bhat-(W1b^lam1-1)/lam1)*((2*W2b^lam2-2)/lam2^3-(2*W2b^lam2*log(W2b))/lam2^2+(W2b^lam2*log(W2b)^2)/lam2))/(s1bhat^2*s2bhat^2))/((2*covb^2)/(s1bhat^2*s2bhat^2)-2)) h1r=c(h1_1,h1_2,h1_3,h1_4,h1_5,h1_6,h1_7,h1_8,h1_9,h1_10,h1_11,h1_12) h2r=c(h2_1,h2_2,h2_3,h2_4,h2_5,h2_6,h2_7,h2_8,h2_9,h2_10,h2_11,h2_12) h3r=c(h3_1,h3_2,h3_3,h3_4,h3_5,h3_6,h3_7,h3_8,h3_9,h3_10,h3_11,h3_12) h4r=c(h4_1,h4_2,h4_3,h4_4,h4_5,h4_6,h4_7,h4_8,h4_9,h4_10,h4_11,h4_12) h5r=c(h5_1,h5_2,h5_3,h5_4,h5_5,h5_6,h5_7,h5_8,h5_9,h5_10,h5_11,h5_12) h6r=c(h6_1,h6_2,h6_3,h6_4,h6_5,h6_6,h6_7,h6_8,h6_9,h6_10,h6_11,h6_12) h7r=c(h7_1,h7_2,h7_3,h7_4,h7_5,h7_6,h7_7,h7_8,h7_9,h7_10,h7_11,h7_12) h8r=c(h8_1,h8_2,h8_3,h8_4,h8_5,h8_6,h8_7,h8_8,h8_9,h8_10,h8_11,h8_12) h9r=c(h9_1,h9_2,h9_3,h9_4,h9_5,h9_6,h9_7,h9_8,h9_9,h9_10,h9_11,h9_12) h10r=c(h10_1,h10_2,h10_3,h10_4,h10_5,h10_6,h10_7,h10_8,h10_9,h10_10,h10_11,h10_12) h11r=c(h11_1,h11_2,h11_3,h11_4,h11_5,h11_6,h11_7,h11_8,h11_9,h11_10,h11_11,h11_12) h12r=c(h12_1,h12_2,h12_3,h12_4,h12_5,h12_6,h12_7,h12_8,h12_9,h12_10,h12_11,h12_12) HCF=rbind(h1r,h2r,h3r,h4r,h5r,h6r,h7r,h8r,h9r,h10r,h11r,h12r) HCF[1,2]=0;HCF[1,6]=0;HCF[1,9]=0; HCF[2,1]=0;HCF[2,5]=0; HCF[3,4]=0;HCF[3,8]=0;HCF[3,10]=0;HCF[4,3]=0; HCF[4,7]=0; HCF[5,2]=0;HCF[5,6]=0;HCF[5,9]=0; HCF[6,1]=0;HCF[6,5]=0; HCF[7,4]=0;HCF[7,8]=0;HCF[7,10]=0; HCF[8,3]=0;HCF[8,7]=0; HCF[9,1]=0;HCF[9,5]=0; HCF[10,3]=0; HCF[10,7]=0; HCF[1,1]=na*HCF[1,1]; HCF[1,5]=na*HCF[1,5]; HCF[5,1]=na*HCF[5,1]; HCF[3,3]=nb*HCF[3,3]; HCF[3,7]=nb*HCF[3,7]; HCF[7,3]=nb*HCF[7,3]; HCF[7,7]=nb*HCF[7,7]; HCF[5,5]=na*HCF[5,5]; VV=inv(-HCF) ############################################################################# ############################################################################# #================ DIFFERENCES OF THE YOUDEN INDICES======================= m1a=m1ahat; s1a=s1ahat; m1b=m1bhat; s1b=s1bhat; m2a=m2ahat; s2a=s2ahat; m2b=m2bhat; s2b=s2bhat; a1=m1b-m1a; b1=s1b/s1a; c1=(m1a*(b1^2-1)-a1+b1*sqrt(a1^2+(b1^2-1)*s1a^2*log(b1^2)))/(b1^2-1); J1=pnorm((m1b-c1)/s1b)+pnorm((c1-m1a)/s1a)-1 J1BCCC=J1 a2=m2b-m2a; b2=s2b/s2a; c2=(m2a*(b2^2-1)-a2+b2*sqrt(a2^2+(b2^2-1)*s2a^2*log(b2^2)))/(b2^2-1); J2=pnorm((m2b-c2)/s2b)+pnorm((c2-m2a)/s2a)-1 J2BCCC=J2 dJ1_dm1a = (2^(1/2)*exp(-(m1a - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1a^2))*((s1b^2/s1a^2 + (s1b*(2*m1a - 2*m1b))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)))/(s1b^2/s1a^2 - 1) - 1))/(2*pi^(1/2)*s1a) - (2^(1/2)*exp(-(m1b - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1b^2))*(s1b^2/s1a^2 + (s1b*(2*m1a - 2*m1b))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))))/(2*pi^(1/2)*s1b*(s1b^2/s1a^2 - 1)); dJ1_ds1a = (2^(1/2)*exp(-(m1b - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1b^2))*(((s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a^2 + (2*m1a*s1b^2)/s1a^3 + (s1b*(2*s1a*(s1b^2/s1a^2 - 1) + (2*s1b^2*log(s1b^2/s1a^2))/s1a - 2*s1a*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1)))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)))/(s1b^2/s1a^2 - 1) - (2*s1b^2*(m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a))/(s1a^3*(s1b^2/s1a^2 - 1)^2)))/(2*pi^(1/2)*s1b) - (exp(-(m1a - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1a^2))*((2^(1/2)*(((s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a^2 + (2*m1a*s1b^2)/s1a^3 + (s1b*(2*s1a*(s1b^2/s1a^2 - 1) + (2*s1b^2*log(s1b^2/s1a^2))/s1a - 2*s1a*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1)))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)))/(s1b^2/s1a^2 - 1) - (2*s1b^2*(m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a))/(s1a^3*(s1b^2/s1a^2 - 1)^2)))/(2*s1a) - (2^(1/2)*(m1a - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1)))/(2*s1a^2)))/pi^(1/2); dJ1_dm1b = (2^(1/2)*exp(-(m1b - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1b^2))*(((s1b*(2*m1a - 2*m1b))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)) + 1)/(s1b^2/s1a^2 - 1) + 1))/(2*pi^(1/2)*s1b) - (2^(1/2)*exp(-(m1a - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1a^2))*((s1b*(2*m1a - 2*m1b))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)) + 1))/(2*pi^(1/2)*s1a*(s1b^2/s1a^2 - 1)); dJ1_ds1b = (2^(1/2)*exp(-(m1a - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1a^2))*((((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)/s1a + (2*m1a*s1b)/s1a^2 + (s1b*(2*s1b*log(s1b^2/s1a^2) + (2*s1a^2*(s1b^2/s1a^2 - 1))/s1b))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)))/(s1b^2/s1a^2 - 1) - (2*s1b*(m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a))/(s1a^2*(s1b^2/s1a^2 - 1)^2)))/(2*pi^(1/2)*s1a) - (exp(-(m1b - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1))^2/(2*s1b^2))*((2^(1/2)*((((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)/s1a + (2*m1a*s1b)/s1a^2 + (s1b*(2*s1b*log(s1b^2/s1a^2) + (2*s1a^2*(s1b^2/s1a^2 - 1))/s1b))/(2*s1a*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2)))/(s1b^2/s1a^2 - 1) - (2*s1b*(m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a))/(s1a^2*(s1b^2/s1a^2 - 1)^2)))/(2*s1b) + (2^(1/2)*(m1b - (m1a - m1b + m1a*(s1b^2/s1a^2 - 1) + (s1b*((m1a - m1b)^2 + s1a^2*log(s1b^2/s1a^2)*(s1b^2/s1a^2 - 1))^(1/2))/s1a)/(s1b^2/s1a^2 - 1)))/(2*s1b^2)))/pi^(1/2); dJ1_dm2a =0; dJ1_ds2a =0; dJ1_dm2b =0; dJ1_ds2b =0; dJ2_dm1a =0; dJ2_ds1a =0; dJ2_dm1b =0; dJ2_ds1b =0; dJ2_dm2a = (2^(1/2)*exp(-(m2a - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2a^2))*((s2b^2/s2a^2 + (s2b*(2*m2a - 2*m2b))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)))/(s2b^2/s2a^2 - 1) - 1))/(2*pi^(1/2)*s2a) - (2^(1/2)*exp(-(m2b - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2b^2))*(s2b^2/s2a^2 + (s2b*(2*m2a - 2*m2b))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))))/(2*pi^(1/2)*s2b*(s2b^2/s2a^2 - 1)); dJ2_ds2a = (2^(1/2)*exp(-(m2b - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2b^2))*(((s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a^2 + (2*m2a*s2b^2)/s2a^3 + (s2b*(2*s2a*(s2b^2/s2a^2 - 1) + (2*s2b^2*log(s2b^2/s2a^2))/s2a - 2*s2a*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1)))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)))/(s2b^2/s2a^2 - 1) - (2*s2b^2*(m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a))/(s2a^3*(s2b^2/s2a^2 - 1)^2)))/(2*pi^(1/2)*s2b) - (exp(-(m2a - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2a^2))*((2^(1/2)*(((s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a^2 + (2*m2a*s2b^2)/s2a^3 + (s2b*(2*s2a*(s2b^2/s2a^2 - 1) + (2*s2b^2*log(s2b^2/s2a^2))/s2a - 2*s2a*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1)))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)))/(s2b^2/s2a^2 - 1) - (2*s2b^2*(m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a))/(s2a^3*(s2b^2/s2a^2 - 1)^2)))/(2*s2a) - (2^(1/2)*(m2a - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1)))/(2*s2a^2)))/pi^(1/2); dJ2_dm2b = (2^(1/2)*exp(-(m2b - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2b^2))*(((s2b*(2*m2a - 2*m2b))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)) + 1)/(s2b^2/s2a^2 - 1) + 1))/(2*pi^(1/2)*s2b) - (2^(1/2)*exp(-(m2a - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2a^2))*((s2b*(2*m2a - 2*m2b))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)) + 1))/(2*pi^(1/2)*s2a*(s2b^2/s2a^2 - 1)); dJ2_ds2b = (2^(1/2)*exp(-(m2a - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2a^2))*((((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)/s2a + (2*m2a*s2b)/s2a^2 + (s2b*(2*s2b*log(s2b^2/s2a^2) + (2*s2a^2*(s2b^2/s2a^2 - 1))/s2b))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)))/(s2b^2/s2a^2 - 1) - (2*s2b*(m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a))/(s2a^2*(s2b^2/s2a^2 - 1)^2)))/(2*pi^(1/2)*s2a) - (exp(-(m2b - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1))^2/(2*s2b^2))*((2^(1/2)*((((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)/s2a + (2*m2a*s2b)/s2a^2 + (s2b*(2*s2b*log(s2b^2/s2a^2) + (2*s2a^2*(s2b^2/s2a^2 - 1))/s2b))/(2*s2a*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2)))/(s2b^2/s2a^2 - 1) - (2*s2b*(m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a))/(s2a^2*(s2b^2/s2a^2 - 1)^2)))/(2*s2b) + (2^(1/2)*(m2b - (m2a - m2b + m2a*(s2b^2/s2a^2 - 1) + (s2b*((m2a - m2b)^2 + s2a^2*log(s2b^2/s2a^2)*(s2b^2/s2a^2 - 1))^(1/2))/s2a)/(s2b^2/s2a^2 - 1)))/(2*s2b^2)))/pi^(1/2); VVpars=VV[1:8,1:8]; VJ1=c(dJ1_dm1a, dJ1_ds1a, dJ1_dm1b, dJ1_ds1b, dJ1_dm2a, dJ1_ds2a, dJ1_dm2b, dJ1_ds2b)%*%VVpars%*%t(t(c(dJ1_dm1a, dJ1_ds1a, dJ1_dm1b, dJ1_ds1b, dJ1_dm2a, dJ1_ds2a, dJ1_dm2b, dJ1_ds2b))); VJ2=c(dJ2_dm1a, dJ2_ds1a, dJ2_dm1b, dJ2_ds1b, dJ2_dm2a, dJ2_ds2a, dJ2_dm2b, dJ2_ds2b)%*%VVpars%*%t(t(c(dJ2_dm1a, dJ2_ds1a, dJ2_dm1b, dJ2_ds1b, dJ2_dm2a, dJ2_ds2a, dJ2_dm2b, dJ2_ds2b))); COVJ12=c(dJ1_dm1a, dJ1_ds1a, dJ1_dm1b, dJ1_ds1b, dJ1_dm2a, dJ1_ds2a, dJ1_dm2b, dJ1_ds2b)%*%VVpars%*%t(t(c(dJ2_dm1a, dJ2_ds1a, dJ2_dm1b, dJ2_ds1b, dJ2_dm2a, dJ2_ds2a, dJ2_dm2b, dJ2_ds2b))); kk1=c(VJ1, COVJ12) kk2=c(COVJ12, VJ2) V12=rbind(kk1,kk2) Z=(J2-J1)/(sqrt(VJ1+VJ2-2*COVJ12)); J1original=J1; J2original=J2; SE=sqrt(VJ1+VJ2-2*COVJ12); CIoriginal=c((J2-J1)-pmalpha*SE, (J2-J1)+pmalpha*SE); pval2tJ=2*pnorm(-abs(Z)) #JT2=log(0.5*(J2+1)/(1-0.5*(J2+1))); #JT1=log(0.5*(J1+1)/(1-0.5*(J1+1))); #VJT1=4/(J1^2-1)^2*VJ1; #VJT2=4/(J2^2-1)^2*VJ2; #COVJT12=c(-2/(J1^2-1), 0)%*%V12%*%t(t(c(0, -2/(J2^2-1)))) #Zstar=(JT2-JT1)/(sqrt(VJT1+VJT2-2*COVJT12)) #SE=(sqrt(VJT1+VJT2-2*COVJT12)) #CIZstar=c((JT2-JT1)-pmalpha*SE, (JT2-JT1)+pmalpha*SE) #CIZstar JT2=qnorm(J2); JT1=qnorm(J1); VJT1=(1/(dnorm(qnorm(J1))))^2*VJ1; VJT2=(1/(dnorm(qnorm(J2))))^2*VJ2; COVJT12=c((1/(dnorm(qnorm(J1))))*VJ1, 0)%*%V12%*%t(t(c(0, (1/(dnorm(qnorm(J2))))*VJ2))) Zstar=(JT2-JT1)/(sqrt(VJT1+VJT2-2*COVJT12)) SE=(sqrt(VJT1+VJT2-2*COVJT12)) CIZstar=c((JT2-JT1)-pmalpha*SE, (JT2-JT1)+pmalpha*SE) CIZstar pval2t=2*pnorm(-abs(Zstar)) #CIoriginal=c(2*(exp(CIZstar[1])/(1+exp(CIZstar[1])))-1 , 2*(exp(CIZstar[2])/(1+exp(CIZstar[2])))-1) #====TWO ROC FUNCTIONS: ONE IS THE BOXCOX AND THE OTHER THE REFERENCE LINE================ roc1<-function(t){ na = length(W1alam) nb = length(W1blam) s1alam=sqrt( var(W1alam)*(na-1)/na ) s1blam=sqrt( var(W1blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W1alam), sd=s1alam), mean=mean(W1blam), sd=s1blam) } roc2<-function(t){ na = length(W2alam) nb = length(W2blam) s2alam=sqrt( var(W2alam)*(na-1)/na ) s2blam=sqrt( var(W2blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W2alam), sd=s2alam), mean=mean(W2blam), sd=s2blam) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } Sens1=1-pnorm(c1, mean=mean(W1blam), sd=sqrt(var(W1blam)*(nb-1)/nb)) Spec1= pnorm(c1, mean=mean(W1alam), sd=sqrt(var(W1alam)*(na-1)/na)) Sens2=1-pnorm(c2, mean=mean(W2blam), sd=sqrt(var(W2blam)*(nb-1)/nb)) Spec2= pnorm(c2, mean=mean(W2alam), sd=sqrt(var(W2alam)*(na-1)/na)) #================IF PLOTS ARE REQUESTED PLOT THE ROCS== if (plots=="on") { # x11() plot(linspace(0,1,1000),roc1(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="red") lines(linspace(0,1,1000),roc2(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="black") lines(linspace(0,1,1000),linspace(0,1,1000),type="l", lty=2) points(1-Spec1,Sens1, col = "red") points(1-Spec2,Sens2, col = "black") #line for the Youden: lines(c(1-Spec1,1-Spec1),c(rocuseless(1-Spec1),Sens1),col="blue") lines(c(1-Spec2,1-Spec2),c(rocuseless(1-Spec2),Sens2),col="green") legend("bottomright", legend=c(paste("ROC for Marker 1 with J =", formattable(J1original, digits = 4, format = "f"), " "), paste("ROC for Marker 2 with J =", formattable(J2original, digits = 4, format = "f"), " "), paste("P-value for the difference:", formattable(pval2t, digits = 4, format = "f"), " ")), col=c("red", "black", "white"), lty=c(1, 1, NA), pch = c(NA, NA, NA), cex=0.8) } # if (plots=="on"){ # points(1-Spec1,Sens1, col = "red") # points(1-Spec2,Sens2, col = "black") # # #line for the Youden: # lines(c(1-Spec1,1-Spec1),c(rocuseless(1-Spec1),Sens1),col="blue") # lines(c(1-Spec2,1-Spec2),c(rocuseless(1-Spec2),Sens2),col="green") # # } #====================================================== res <- data.frame(J1 = formattable(J1original, digits = 4, format = "f"), J2 = formattable(J2original, digits = 4, format = "f"), diff = formattable(J2original - J1original, digits = 4, format = "f"), p_value_probit = formattable(pval2t, digits = 4, format = "f"), p_value = formattable(pval2tJ, digits = 4, format = "f"), ci_ll = formattable(CIoriginal[1], digits = 4, format = "f"), ci_ul = formattable(CIoriginal[2], digits = 4, format = "f")) rownames(res) <- c("Estimates:") colnames(res) <- c("J 1", "J 2", "Diff", "P-Val (Probit)", "P-Val", "CI (LL)", "CI (UL)") res <- formattable(as.matrix(res), digits = 4, format = "f") res #return(list(AUCmarker1=AUC1original,AUCmarker2=AUC2original, pvalue_difference= pval2t, CI_difference= pnorm(CIdiff), rocbc1=roc1, rocbc2=roc2)) #return(list(resultstable=res,J1=J1original,J2=J2original, pvalue_difference= pval2t, CI_difference= CIoriginal, rocbc1=roc1, rocbc2=roc2)) return(list(resultstable=res, J1=J1original, J2=J2original, pvalue_probit_difference= pval2t, pvalue_difference= pval2tJ, CI_difference= CIoriginal, roc1=roc1, roc2=roc2, transx1=W1alam, transy1=W1blam, transformation.parameter.1 = lam[1], transx2=W2alam, transy2=W2blam, transformation.parameter.2 = lam[2])) } } ``` ```{r, echo=FALSE, eval=TRUE} comparebcSens <-function (marker1, marker2, D, atSpec, alpha, plots){ if ((length(marker1) != length(D)) | (length(marker2) != length(D))) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (alpha <= 0 | alpha >= 1) { stop("ERROR: The level of significance, alpha, should be set between 0 and 1. A common choice is 0.05.") } else if (sum(is.na(marker1)) > 0 | sum(is.na(marker2)) > 0 | sum(is.na(D)) > 0 | sum(is.na(atSpec)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if ((!is.numeric(atSpec)) | (length(atSpec) != 1)) { stop("ERROR: 'atSpec' must be a single numeric value.") } else if ((sum(marker1 < 0) > 0) | (sum(marker2 < 2)) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { erf <- function (x) 2 * pnorm(x * sqrt(2)) - 1 erfc <- function (x) 2 * pnorm(x * sqrt(2), lower.tail = FALSE) erfinv <- function (x) qnorm((1 + x)/2)/sqrt(2) erfcinv <- function (x) qnorm(x/2, lower.tail = FALSE)/sqrt(2) pmalpha=qnorm(1-alpha/2); if (plots!="on"){plots="off"} W1a=marker1[D==0] W1b=marker1[D==1] W2a=marker2[D==0] W2b=marker2[D==1] Wa=cbind(W1a,W2a); Wb=cbind(W1b,W2b); na=length(W1a) nb=length(W1b) likbox2D<-function(W1a,W1b,W2a,W2b,lam){ na=length(W1a); nb=length(W1b); out=c(); #for (i in 1:length(h)){ W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) cova= cov(Walam)*(na-1)/na; covb= cov(Wblam)*(nb-1)/nb; mualam=c(mean(W1alam), mean(W2alam)) mublam=c(mean(W1blam), mean(W2blam)) out= -sum(log(dmvnorm(Walam,mualam,cova)))-sum(log(dmvnorm(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b)); return(out) } lam=c(2,2) likbox2D(W1a,W1b,W2a,W2b,lam) logL<-function(h){ likbox2D(W1a,W1b,W2a,W2b,h) } # init=c(0.8,0.8) # lam=fminsearch(logL,init) # lam=c(lam$optbase$xopt) lam<- optim(c(1,1),logL,gr=NULL,method="BFGS", control=list(maxit=10000)) lam=c(lam$par) lam out <- optim(c(1,1), logL, method = "Nelder-Mead") lam = out$par ################################### W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) m1ahat=mean(W1alam) m1bhat=mean(W1blam) m2ahat=mean(W2alam) m2bhat=mean(W2blam) s1ahat=sqrt( var(W1alam)*(na-1)/na ) s1bhat=sqrt( var(W1blam)*(nb-1)/nb ) s2ahat=sqrt( var(W2alam)*(na-1)/na ) s2bhat=sqrt( var(W2blam)*(nb-1)/nb ) cova= cov(Walam)*(na-1)/na;cova=cova[1,2]; covb= cov(Wblam)*(nb-1)/nb;covb=covb[1,2]; lam1=lam[1] lam2=lam[2] ################################## h1_1=sum(-2/(2*s1ahat^2 - (2*cova^2)/s2ahat^2)) h1_2=sum(-(2*s1ahat*s2ahat^2*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_3=0 h1_4=0; h1_5=sum(cova/(- cova^2 + s1ahat^2*s2ahat^2)); h1_6=sum((s2ahat*(2*cova^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat) - 2*cova*s1ahat^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_7=0 h1_8=0 h1_9=sum((cova^2*(lam2*m2ahat - W2a^lam2 + 1) + s2ahat^2*((lam2*m2ahat - W2a^lam2 + 1)*s1ahat^2 - 2*cova*lam2*m1ahat))/(lam2*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s2ahat^2*(2*lam2 - 2*W1a^lam1*lam2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h1_10=0 h1_11=sum((2*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h1_12=sum(-(cova*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- cova^2 + s1ahat^2*s2ahat^2))) h2_1=sum(-(2*s1ahat*s2ahat^2*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_2=sum(- (3*s1ahat^2*s2ahat^6 + cova^4*(lam1^2*m2ahat^2 + lam1^2*s2ahat^2) - cova^3*(2*lam1*m2ahat*s2ahat^2 - 2*W1a^lam1*lam1*m2ahat*s2ahat^2 + 2*lam1^2*m1ahat*m2ahat*s2ahat^2) - cova*(6*lam1*m2ahat*s1ahat^2*s2ahat^4 - 6*W1a^lam1*lam1*m2ahat*s1ahat^2*s2ahat^4 + 6*lam1^2*m1ahat*m2ahat*s1ahat^2*s2ahat^4) + cova^2*(W1a^(2*lam1)*s2ahat^4 - 2*W1a^lam1*s2ahat^4 + s2ahat^4 + 2*lam1*m1ahat*s2ahat^4 + lam1^2*m1ahat^2*s2ahat^4 - 2*W1a^lam1*lam1*m1ahat*s2ahat^4 + 3*lam1^2*m2ahat^2*s1ahat^2*s2ahat^2) - 6*W1a^lam1*s1ahat^2*s2ahat^6 + 3*W1a^(2*lam1)*s1ahat^2*s2ahat^6 - lam1^2*s1ahat^4*s2ahat^6 + 3*lam1^2*m1ahat^2*s1ahat^2*s2ahat^6 + 6*lam1*m1ahat*s1ahat^2*s2ahat^6 - 6*W1a^lam1*lam1*m1ahat*s1ahat^2*s2ahat^6)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^4*(W2a^(2*lam2)*lam1^2 - 2*W2a^lam2*lam1^2 + lam1^2) + cova^2*(3*lam1^2*s1ahat^2*s2ahat^2 + 3*W2a^(2*lam2)*lam1^2*s1ahat^2*s2ahat^2 - 6*W2a^lam2*lam1^2*s1ahat^2*s2ahat^2) - lam2*((2*W2a^lam2*lam1^2*m2ahat - 2*lam1^2*m2ahat)*cova^4 + (2*lam1*s2ahat^2 + 2*lam1^2*m1ahat*s2ahat^2 - 2*W1a^lam1*lam1*s2ahat^2 - 2*W2a^lam2*lam1*s2ahat^2 - 2*W2a^lam2*lam1^2*m1ahat*s2ahat^2 + 2*W1a^lam1*W2a^lam2*lam1*s2ahat^2)*cova^3 + (6*W2a^lam2*lam1^2*m2ahat*s1ahat^2*s2ahat^2 - 6*lam1^2*m2ahat*s1ahat^2*s2ahat^2)*cova^2 + (6*lam1*s1ahat^2*s2ahat^4 + 6*lam1^2*m1ahat*s1ahat^2*s2ahat^4 - 6*W1a^lam1*lam1*s1ahat^2*s2ahat^4 - 6*W2a^lam2*lam1*s1ahat^2*s2ahat^4 - 6*W2a^lam2*lam1^2*m1ahat*s1ahat^2*s2ahat^4 + 6*W1a^lam1*W2a^lam2*lam1*s1ahat^2*s2ahat^4)*cova))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_3=sum(0) h2_4=sum(0) h2_5=sum((s1ahat*(2*cova^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat) - 2*cova*s2ahat^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_6=sum((lam2*((2*cova*s2ahat^3*(2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat) - 2*cova*s2ahat*(4*cova*lam1^2*m2ahat - 4*W2a^lam2*cova*lam1^2*m2ahat))*s1ahat^3 + 2*cova*s2ahat*(2*cova^2*lam1 + 2*cova^2*lam1^2*m1ahat - 2*W1a^lam1*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1^2*m1ahat + 2*W1a^lam1*W2a^lam2*cova^2*lam1)*s1ahat) - 2*cova*s1ahat^3*s2ahat*(2*cova*lam1^2 + 2*W2a^(2*lam2)*cova*lam1^2 - 4*W2a^lam2*cova*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - ((4*cova^2*lam1^2*m2ahat^2*s2ahat - 2*cova*s2ahat^3*(2*lam1*m2ahat + cova*lam1^2 + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat))*s1ahat^3 + (2*cova*(2*cova + 2*W1a^(2*lam1)*cova - 4*W1a^lam1*cova + 2*cova*lam1^2*m1ahat^2 + 4*cova*lam1*m1ahat - 4*W1a^lam1*cova*lam1*m1ahat)*s2ahat^3 + 2*cova*(cova^3*lam1^2 - 2*cova^2*lam1*m2ahat + 2*W1a^lam1*cova^2*lam1*m2ahat - 2*cova^2*lam1^2*m1ahat*m2ahat)*s2ahat)*s1ahat)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_7=sum(0) h2_8=sum(0) h2_9=sum(- (cova*(s2ahat^4*(- 4*lam2^2*m1ahat^2*s1ahat + 2*lam2^2*s1ahat^3) - s1ahat^3*s2ahat^2*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2)) - cova^3*(s1ahat*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2) + 2*lam2^2*s1ahat*s2ahat^2) + s1ahat^3*s2ahat^4*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + cova^2*s1ahat*s2ahat^2*(6*lam2*m1ahat + 6*lam2^2*m1ahat*m2ahat - 6*W2a^lam2*lam2*m1ahat))/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2ahat - 6*W1a^lam1*lam2 - 6*W2a^lam2*lam2 + 6*W1a^lam1*W2a^lam2*lam2 - 6*W1a^lam1*lam2^2*m2ahat)*cova^2*s1ahat*s2ahat^2 + (8*W1a^lam1*lam2^2*m1ahat - 8*lam2^2*m1ahat)*cova*s1ahat*s2ahat^4 + (2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat)*s1ahat^3*s2ahat^4) - cova*s1ahat*s2ahat^4*(4*W1a^(2*lam1)*lam2^2 - 8*W1a^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h2_10=sum(0) h2_11=sum((2*s1ahat*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h2_12=sum(-(2*cova*s1ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h3_1=sum(0) h3_2=sum(0) h3_3=sum(-2/(2*s1bhat^2 - (2*covb^2)/s2bhat^2)) h3_4=sum(-(2*s1bhat*s2bhat^2*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_5=sum(0) h3_6=sum(0) h3_7=sum(covb/(- covb^2 + s1bhat^2*s2bhat^2)) h3_8=sum((s2bhat*(2*covb^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat) - 2*covb*s1bhat^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_9=sum(0) h3_10=sum((covb^2*(lam2*m2bhat - W2b^lam2 + 1) + s2bhat^2*((lam2*m2bhat - W2b^lam2 + 1)*s1bhat^2 - 2*covb*lam2*m1bhat))/(lam2*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s2bhat^2*(2*lam2 - 2*W1b^lam1*lam2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h3_11=sum((2*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h3_12=sum(-(covb*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- covb^2 + s1bhat^2*s2bhat^2))) h4_1=sum(0) h4_2=sum(0) h4_3=sum(-(2*s1bhat*s2bhat^2*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_4=sum(- (3*s1bhat^2*s2bhat^6 + covb^4*(lam1^2*m2bhat^2 + lam1^2*s2bhat^2) - covb^3*(2*lam1*m2bhat*s2bhat^2 - 2*W1b^lam1*lam1*m2bhat*s2bhat^2 + 2*lam1^2*m1bhat*m2bhat*s2bhat^2) - covb*(6*lam1*m2bhat*s1bhat^2*s2bhat^4 - 6*W1b^lam1*lam1*m2bhat*s1bhat^2*s2bhat^4 + 6*lam1^2*m1bhat*m2bhat*s1bhat^2*s2bhat^4) + covb^2*(W1b^(2*lam1)*s2bhat^4 - 2*W1b^lam1*s2bhat^4 + s2bhat^4 + 2*lam1*m1bhat*s2bhat^4 + lam1^2*m1bhat^2*s2bhat^4 - 2*W1b^lam1*lam1*m1bhat*s2bhat^4 + 3*lam1^2*m2bhat^2*s1bhat^2*s2bhat^2) - 6*W1b^lam1*s1bhat^2*s2bhat^6 + 3*W1b^(2*lam1)*s1bhat^2*s2bhat^6 - lam1^2*s1bhat^4*s2bhat^6 + 3*lam1^2*m1bhat^2*s1bhat^2*s2bhat^6 + 6*lam1*m1bhat*s1bhat^2*s2bhat^6 - 6*W1b^lam1*lam1*m1bhat*s1bhat^2*s2bhat^6)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^4*(W2b^(2*lam2)*lam1^2 - 2*W2b^lam2*lam1^2 + lam1^2) + covb^2*(3*lam1^2*s1bhat^2*s2bhat^2 + 3*W2b^(2*lam2)*lam1^2*s1bhat^2*s2bhat^2 - 6*W2b^lam2*lam1^2*s1bhat^2*s2bhat^2) - lam2*((2*W2b^lam2*lam1^2*m2bhat - 2*lam1^2*m2bhat)*covb^4 + (2*lam1*s2bhat^2 + 2*lam1^2*m1bhat*s2bhat^2 - 2*W1b^lam1*lam1*s2bhat^2 - 2*W2b^lam2*lam1*s2bhat^2 - 2*W2b^lam2*lam1^2*m1bhat*s2bhat^2 + 2*W1b^lam1*W2b^lam2*lam1*s2bhat^2)*covb^3 + (6*W2b^lam2*lam1^2*m2bhat*s1bhat^2*s2bhat^2 - 6*lam1^2*m2bhat*s1bhat^2*s2bhat^2)*covb^2 + (6*lam1*s1bhat^2*s2bhat^4 + 6*lam1^2*m1bhat*s1bhat^2*s2bhat^4 - 6*W1b^lam1*lam1*s1bhat^2*s2bhat^4 - 6*W2b^lam2*lam1*s1bhat^2*s2bhat^4 - 6*W2b^lam2*lam1^2*m1bhat*s1bhat^2*s2bhat^4 + 6*W1b^lam1*W2b^lam2*lam1*s1bhat^2*s2bhat^4)*covb))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_5=sum(0) h4_6=sum(0) h4_7=sum((s1bhat*(2*covb^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat) - 2*covb*s2bhat^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_8=sum((lam2*((2*covb*s2bhat^3*(2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat) - 2*covb*s2bhat*(4*covb*lam1^2*m2bhat - 4*W2b^lam2*covb*lam1^2*m2bhat))*s1bhat^3 + 2*covb*s2bhat*(2*covb^2*lam1 + 2*covb^2*lam1^2*m1bhat - 2*W1b^lam1*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1^2*m1bhat + 2*W1b^lam1*W2b^lam2*covb^2*lam1)*s1bhat) - 2*covb*s1bhat^3*s2bhat*(2*covb*lam1^2 + 2*W2b^(2*lam2)*covb*lam1^2 - 4*W2b^lam2*covb*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - ((4*covb^2*lam1^2*m2bhat^2*s2bhat - 2*covb*s2bhat^3*(2*lam1*m2bhat + covb*lam1^2 + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat))*s1bhat^3 + (2*covb*(2*covb + 2*W1b^(2*lam1)*covb - 4*W1b^lam1*covb + 2*covb*lam1^2*m1bhat^2 + 4*covb*lam1*m1bhat - 4*W1b^lam1*covb*lam1*m1bhat)*s2bhat^3 + 2*covb*(covb^3*lam1^2 - 2*covb^2*lam1*m2bhat + 2*W1b^lam1*covb^2*lam1*m2bhat - 2*covb^2*lam1^2*m1bhat*m2bhat)*s2bhat)*s1bhat)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_9=sum(0) h4_10=sum(- (covb*(s2bhat^4*(- 4*lam2^2*m1bhat^2*s1bhat + 2*lam2^2*s1bhat^3) - s1bhat^3*s2bhat^2*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2)) - covb^3*(s1bhat*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2) + 2*lam2^2*s1bhat*s2bhat^2) + s1bhat^3*s2bhat^4*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + covb^2*s1bhat*s2bhat^2*(6*lam2*m1bhat + 6*lam2^2*m1bhat*m2bhat - 6*W2b^lam2*lam2*m1bhat))/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2bhat - 6*W1b^lam1*lam2 - 6*W2b^lam2*lam2 + 6*W1b^lam1*W2b^lam2*lam2 - 6*W1b^lam1*lam2^2*m2bhat)*covb^2*s1bhat*s2bhat^2 + (8*W1b^lam1*lam2^2*m1bhat - 8*lam2^2*m1bhat)*covb*s1bhat*s2bhat^4 + (2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat)*s1bhat^3*s2bhat^4) - covb*s1bhat*s2bhat^4*(4*W1b^(2*lam1)*lam2^2 - 8*W1b^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h4_11=sum((2*s1bhat*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h4_12=sum(-(2*covb*s1bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h5_1=sum(cova/(- cova^2 + s1ahat^2*s2ahat^2)) h5_2=sum((s1ahat*(2*cova^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat) - 2*cova*s2ahat^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_3=0 h5_4=0 h5_5=sum(-2/(2*s2ahat^2 - (2*cova^2)/s1ahat^2)) h5_6=sum(-(2*s1ahat^2*s2ahat*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_7=0 h5_8=0 h5_9=sum((cova^2*(lam1*m1ahat - W1a^lam1 + 1) + s1ahat^2*((lam1*m1ahat - W1a^lam1 + 1)*s2ahat^2 - 2*cova*lam1*m2ahat))/(lam1*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s1ahat^2*(2*lam1 - 2*W2a^lam2*lam1))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h5_10=0 h5_11=sum(-(cova*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- cova^2 + s1ahat^2*s2ahat^2))) h5_12=sum((2*s1ahat^2*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h6_1=sum((s2ahat*(2*cova^2*(lam2 - W1a^lam1*lam2 + lam1*lam2*m1ahat) - 2*cova*s1ahat^2*(lam1 - W2a^lam2*lam1 + lam1*lam2*m2ahat)))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_2=sum((lam2*((2*cova*s2ahat^3*(2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat) - 2*cova*s2ahat*(4*cova*lam1^2*m2ahat - 4*W2a^lam2*cova*lam1^2*m2ahat))*s1ahat^3 + 2*cova*s2ahat*(2*cova^2*lam1 + 2*cova^2*lam1^2*m1ahat - 2*W1a^lam1*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1 - 2*W2a^lam2*cova^2*lam1^2*m1ahat + 2*W1a^lam1*W2a^lam2*cova^2*lam1)*s1ahat) - 2*cova*s1ahat^3*s2ahat*(2*cova*lam1^2 + 2*W2a^(2*lam2)*cova*lam1^2 - 4*W2a^lam2*cova*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - ((4*cova^2*lam1^2*m2ahat^2*s2ahat - 2*cova*s2ahat^3*(2*lam1*m2ahat + cova*lam1^2 + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat))*s1ahat^3 + (2*cova*(2*cova + 2*W1a^(2*lam1)*cova - 4*W1a^lam1*cova + 2*cova*lam1^2*m1ahat^2 + 4*cova*lam1*m1ahat - 4*W1a^lam1*cova*lam1*m1ahat)*s2ahat^3 + 2*cova*(cova^3*lam1^2 - 2*cova^2*lam1*m2ahat + 2*W1a^lam1*cova^2*lam1*m2ahat - 2*cova^2*lam1^2*m1ahat*m2ahat)*s2ahat)*s1ahat)/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_3=0 h6_4=0 h6_5=sum(-(2*s1ahat^2*s2ahat*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_6=sum(- (3*s1ahat^6*s2ahat^2 + cova^4*(lam2^2*m1ahat^2 + lam2^2*s1ahat^2) - cova^3*(2*lam2*m1ahat*s1ahat^2 - 2*W2a^lam2*lam2*m1ahat*s1ahat^2 + 2*lam2^2*m1ahat*m2ahat*s1ahat^2) - cova*(6*lam2*m1ahat*s1ahat^4*s2ahat^2 - 6*W2a^lam2*lam2*m1ahat*s1ahat^4*s2ahat^2 + 6*lam2^2*m1ahat*m2ahat*s1ahat^4*s2ahat^2) + cova^2*(W2a^(2*lam2)*s1ahat^4 - 2*W2a^lam2*s1ahat^4 + s1ahat^4 + 2*lam2*m2ahat*s1ahat^4 + lam2^2*m2ahat^2*s1ahat^4 - 2*W2a^lam2*lam2*m2ahat*s1ahat^4 + 3*lam2^2*m1ahat^2*s1ahat^2*s2ahat^2) - 6*W2a^lam2*s1ahat^6*s2ahat^2 + 3*W2a^(2*lam2)*s1ahat^6*s2ahat^2 - lam2^2*s1ahat^6*s2ahat^4 + 3*lam2^2*m2ahat^2*s1ahat^6*s2ahat^2 + 6*lam2*m2ahat*s1ahat^6*s2ahat^2 - 6*W2a^lam2*lam2*m2ahat*s1ahat^6*s2ahat^2)/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^4*(W1a^(2*lam1)*lam2^2 - 2*W1a^lam1*lam2^2 + lam2^2) + cova^2*(3*lam2^2*s1ahat^2*s2ahat^2 + 3*W1a^(2*lam1)*lam2^2*s1ahat^2*s2ahat^2 - 6*W1a^lam1*lam2^2*s1ahat^2*s2ahat^2) - lam1*((2*W1a^lam1*lam2^2*m1ahat - 2*lam2^2*m1ahat)*cova^4 + (2*lam2*s1ahat^2 + 2*lam2^2*m2ahat*s1ahat^2 - 2*W1a^lam1*lam2*s1ahat^2 - 2*W2a^lam2*lam2*s1ahat^2 - 2*W1a^lam1*lam2^2*m2ahat*s1ahat^2 + 2*W1a^lam1*W2a^lam2*lam2*s1ahat^2)*cova^3 + (6*W1a^lam1*lam2^2*m1ahat*s1ahat^2*s2ahat^2 - 6*lam2^2*m1ahat*s1ahat^2*s2ahat^2)*cova^2 + (6*lam2*s1ahat^4*s2ahat^2 + 6*lam2^2*m2ahat*s1ahat^4*s2ahat^2 - 6*W1a^lam1*lam2*s1ahat^4*s2ahat^2 - 6*W2a^lam2*lam2*s1ahat^4*s2ahat^2 - 6*W1a^lam1*lam2^2*m2ahat*s1ahat^4*s2ahat^2 + 6*W1a^lam1*W2a^lam2*lam2*s1ahat^4*s2ahat^2)*cova))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_7=0 h6_8=0 h6_9=sum(- (cova*(s1ahat^4*(- 4*lam1^2*m2ahat^2*s2ahat + 2*lam1^2*s2ahat^3) - s1ahat^2*s2ahat^3*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2)) - cova^3*(s2ahat*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2) + 2*lam1^2*s1ahat^2*s2ahat) + s1ahat^4*s2ahat^3*(2*lam1*m2ahat + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat) + cova^2*s1ahat^2*s2ahat*(6*lam1*m2ahat + 6*lam1^2*m1ahat*m2ahat - 6*W1a^lam1*lam1*m2ahat))/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1ahat - 6*W1a^lam1*lam1 - 6*W2a^lam2*lam1 + 6*W1a^lam1*W2a^lam2*lam1 - 6*W2a^lam2*lam1^2*m1ahat)*cova^2*s1ahat^2*s2ahat + (8*W2a^lam2*lam1^2*m2ahat - 8*lam1^2*m2ahat)*cova*s1ahat^4*s2ahat + (2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat)*s1ahat^4*s2ahat^3) - cova*s1ahat^4*s2ahat*(4*W2a^(2*lam2)*lam1^2 - 8*W2a^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h6_10=0 h6_11=sum(-(2*cova*s2ahat*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h6_12=sum((2*s1ahat^2*s2ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h7_1=0 h7_2=0 h7_3=sum(covb/(- covb^2 + s1bhat^2*s2bhat^2)) h7_4=sum((s1bhat*(2*covb^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat) - 2*covb*s2bhat^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_5=0 h7_6=0 h7_7=sum(-2/(2*s2bhat^2 - (2*covb^2)/s1bhat^2)) h7_8=sum(-(2*s1bhat^2*s2bhat*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_9=0 h7_10=sum((covb^2*(lam1*m1bhat - W1b^lam1 + 1) + s1bhat^2*((lam1*m1bhat - W1b^lam1 + 1)*s2bhat^2 - 2*covb*lam1*m2bhat))/(lam1*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s1bhat^2*(2*lam1 - 2*W2b^lam2*lam1))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h7_11=sum(-(covb*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- covb^2 + s1bhat^2*s2bhat^2))) h7_12=sum((2*s1bhat^2*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h8_1=0 h8_2=0 h8_3=sum((s2bhat*(2*covb^2*(lam2 - W1b^lam1*lam2 + lam1*lam2*m1bhat) - 2*covb*s1bhat^2*(lam1 - W2b^lam2*lam1 + lam1*lam2*m2bhat)))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_4=sum((lam2*((2*covb*s2bhat^3*(2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat) - 2*covb*s2bhat*(4*covb*lam1^2*m2bhat - 4*W2b^lam2*covb*lam1^2*m2bhat))*s1bhat^3 + 2*covb*s2bhat*(2*covb^2*lam1 + 2*covb^2*lam1^2*m1bhat - 2*W1b^lam1*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1 - 2*W2b^lam2*covb^2*lam1^2*m1bhat + 2*W1b^lam1*W2b^lam2*covb^2*lam1)*s1bhat) - 2*covb*s1bhat^3*s2bhat*(2*covb*lam1^2 + 2*W2b^(2*lam2)*covb*lam1^2 - 4*W2b^lam2*covb*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - ((4*covb^2*lam1^2*m2bhat^2*s2bhat - 2*covb*s2bhat^3*(2*lam1*m2bhat + covb*lam1^2 + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat))*s1bhat^3 + (2*covb*(2*covb + 2*W1b^(2*lam1)*covb - 4*W1b^lam1*covb + 2*covb*lam1^2*m1bhat^2 + 4*covb*lam1*m1bhat - 4*W1b^lam1*covb*lam1*m1bhat)*s2bhat^3 + 2*covb*(covb^3*lam1^2 - 2*covb^2*lam1*m2bhat + 2*W1b^lam1*covb^2*lam1*m2bhat - 2*covb^2*lam1^2*m1bhat*m2bhat)*s2bhat)*s1bhat)/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_5=0 h8_6=0 h8_7=sum(-(2*s1bhat^2*s2bhat*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_8=sum(- (3*s1bhat^6*s2bhat^2 + covb^4*(lam2^2*m1bhat^2 + lam2^2*s1bhat^2) - covb^3*(2*lam2*m1bhat*s1bhat^2 - 2*W2b^lam2*lam2*m1bhat*s1bhat^2 + 2*lam2^2*m1bhat*m2bhat*s1bhat^2) - covb*(6*lam2*m1bhat*s1bhat^4*s2bhat^2 - 6*W2b^lam2*lam2*m1bhat*s1bhat^4*s2bhat^2 + 6*lam2^2*m1bhat*m2bhat*s1bhat^4*s2bhat^2) + covb^2*(W2b^(2*lam2)*s1bhat^4 - 2*W2b^lam2*s1bhat^4 + s1bhat^4 + 2*lam2*m2bhat*s1bhat^4 + lam2^2*m2bhat^2*s1bhat^4 - 2*W2b^lam2*lam2*m2bhat*s1bhat^4 + 3*lam2^2*m1bhat^2*s1bhat^2*s2bhat^2) - 6*W2b^lam2*s1bhat^6*s2bhat^2 + 3*W2b^(2*lam2)*s1bhat^6*s2bhat^2 - lam2^2*s1bhat^6*s2bhat^4 + 3*lam2^2*m2bhat^2*s1bhat^6*s2bhat^2 + 6*lam2*m2bhat*s1bhat^6*s2bhat^2 - 6*W2b^lam2*lam2*m2bhat*s1bhat^6*s2bhat^2)/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^4*(W1b^(2*lam1)*lam2^2 - 2*W1b^lam1*lam2^2 + lam2^2) + covb^2*(3*lam2^2*s1bhat^2*s2bhat^2 + 3*W1b^(2*lam1)*lam2^2*s1bhat^2*s2bhat^2 - 6*W1b^lam1*lam2^2*s1bhat^2*s2bhat^2) - lam1*((2*W1b^lam1*lam2^2*m1bhat - 2*lam2^2*m1bhat)*covb^4 + (2*lam2*s1bhat^2 + 2*lam2^2*m2bhat*s1bhat^2 - 2*W1b^lam1*lam2*s1bhat^2 - 2*W2b^lam2*lam2*s1bhat^2 - 2*W1b^lam1*lam2^2*m2bhat*s1bhat^2 + 2*W1b^lam1*W2b^lam2*lam2*s1bhat^2)*covb^3 + (6*W1b^lam1*lam2^2*m1bhat*s1bhat^2*s2bhat^2 - 6*lam2^2*m1bhat*s1bhat^2*s2bhat^2)*covb^2 + (6*lam2*s1bhat^4*s2bhat^2 + 6*lam2^2*m2bhat*s1bhat^4*s2bhat^2 - 6*W1b^lam1*lam2*s1bhat^4*s2bhat^2 - 6*W2b^lam2*lam2*s1bhat^4*s2bhat^2 - 6*W1b^lam1*lam2^2*m2bhat*s1bhat^4*s2bhat^2 + 6*W1b^lam1*W2b^lam2*lam2*s1bhat^4*s2bhat^2)*covb))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_9=0 h8_10=sum(- (covb*(s1bhat^4*(- 4*lam1^2*m2bhat^2*s2bhat + 2*lam1^2*s2bhat^3) - s1bhat^2*s2bhat^3*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2)) - covb^3*(s2bhat*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2) + 2*lam1^2*s1bhat^2*s2bhat) + s1bhat^4*s2bhat^3*(2*lam1*m2bhat + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat) + covb^2*s1bhat^2*s2bhat*(6*lam1*m2bhat + 6*lam1^2*m1bhat*m2bhat - 6*W1b^lam1*lam1*m2bhat))/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1bhat - 6*W1b^lam1*lam1 - 6*W2b^lam2*lam1 + 6*W1b^lam1*W2b^lam2*lam1 - 6*W2b^lam2*lam1^2*m1bhat)*covb^2*s1bhat^2*s2bhat + (8*W2b^lam2*lam1^2*m2bhat - 8*lam1^2*m2bhat)*covb*s1bhat^4*s2bhat + (2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat)*s1bhat^4*s2bhat^3) - covb*s1bhat^4*s2bhat*(4*W2b^(2*lam2)*lam1^2 - 8*W2b^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h8_11=sum(-(2*covb*s2bhat*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h8_12=sum((2*s1bhat^2*s2bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h9_1=sum((cova^2*(lam2*m2ahat - W2a^lam2 + 1) + s2ahat^2*((lam2*m2ahat - W2a^lam2 + 1)*s1ahat^2 - 2*cova*lam2*m1ahat))/(lam2*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s2ahat^2*(2*lam2 - 2*W1a^lam1*lam2))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_2=sum(- (cova*(s2ahat^4*(- 4*lam2^2*m1ahat^2*s1ahat + 2*lam2^2*s1ahat^3) - s1ahat^3*s2ahat^2*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2)) - cova^3*(s1ahat*(4*lam2*m2ahat - 4*W2a^lam2 + 2*W2a^(2*lam2) + 2*lam2^2*m2ahat^2 - 4*W2a^lam2*lam2*m2ahat + 2) + 2*lam2^2*s1ahat*s2ahat^2) + s1ahat^3*s2ahat^4*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + cova^2*s1ahat*s2ahat^2*(6*lam2*m1ahat + 6*lam2^2*m1ahat*m2ahat - 6*W2a^lam2*lam2*m1ahat))/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2ahat - 6*W1a^lam1*lam2 - 6*W2a^lam2*lam2 + 6*W1a^lam1*W2a^lam2*lam2 - 6*W1a^lam1*lam2^2*m2ahat)*cova^2*s1ahat*s2ahat^2 + (8*W1a^lam1*lam2^2*m1ahat - 8*lam2^2*m1ahat)*cova*s1ahat*s2ahat^4 + (2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat)*s1ahat^3*s2ahat^4) - cova*s1ahat*s2ahat^4*(4*W1a^(2*lam1)*lam2^2 - 8*W1a^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_3=0 h9_4=0 h9_5=sum((cova^2*(lam1*m1ahat - W1a^lam1 + 1) + s1ahat^2*((lam1*m1ahat - W1a^lam1 + 1)*s2ahat^2 - 2*cova*lam1*m2ahat))/(lam1*(s1ahat^2*s2ahat^2 - cova^2)^2) - (cova*s1ahat^2*(2*lam1 - 2*W2a^lam2*lam1))/(lam1*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_6=sum(- (cova*(s1ahat^4*(- 4*lam1^2*m2ahat^2*s2ahat + 2*lam1^2*s2ahat^3) - s1ahat^2*s2ahat^3*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2)) - cova^3*(s2ahat*(4*lam1*m1ahat - 4*W1a^lam1 + 2*W1a^(2*lam1) + 2*lam1^2*m1ahat^2 - 4*W1a^lam1*lam1*m1ahat + 2) + 2*lam1^2*s1ahat^2*s2ahat) + s1ahat^4*s2ahat^3*(2*lam1*m2ahat + 2*lam1^2*m1ahat*m2ahat - 2*W1a^lam1*lam1*m2ahat) + cova^2*s1ahat^2*s2ahat*(6*lam1*m2ahat + 6*lam1^2*m1ahat*m2ahat - 6*W1a^lam1*lam1*m2ahat))/(lam1^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1ahat - 6*W1a^lam1*lam1 - 6*W2a^lam2*lam1 + 6*W1a^lam1*W2a^lam2*lam1 - 6*W2a^lam2*lam1^2*m1ahat)*cova^2*s1ahat^2*s2ahat + (8*W2a^lam2*lam1^2*m2ahat - 8*lam1^2*m2ahat)*cova*s1ahat^4*s2ahat + (2*lam1 + 2*lam1^2*m1ahat - 2*W1a^lam1*lam1 - 2*W2a^lam2*lam1 + 2*W1a^lam1*W2a^lam2*lam1 - 2*W2a^lam2*lam1^2*m1ahat)*s1ahat^4*s2ahat^3) - cova*s1ahat^4*s2ahat*(4*W2a^(2*lam2)*lam1^2 - 8*W2a^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_7=0 h9_8=0 h9_9=sum(- (s1ahat^2*(cova^2*(6*lam2*m2ahat - 6*W2a^lam2 + 3*W2a^(2*lam2) + 3*lam2^2*m2ahat^2 - 6*W2a^lam2*lam2*m2ahat + 3) - cova*(6*lam2*m1ahat*s2ahat^2 - 6*W2a^lam2*lam2*m1ahat*s2ahat^2 + 6*lam2^2*m1ahat*m2ahat*s2ahat^2) + lam2^2*m1ahat^2*s2ahat^4) - cova^3*(2*lam2*m1ahat + 2*lam2^2*m1ahat*m2ahat - 2*W2a^lam2*lam2*m1ahat) + s1ahat^4*(W2a^(2*lam2)*s2ahat^2 - 2*W2a^lam2*s2ahat^2 + s2ahat^2 - lam2^2*s2ahat^4 + 2*lam2*m2ahat*s2ahat^2 + lam2^2*m2ahat^2*s2ahat^2 - 2*W2a^lam2*lam2*m2ahat*s2ahat^2) + cova^4*lam2^2 + 3*cova^2*lam2^2*m1ahat^2*s2ahat^2)/(lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3) - (cova^2*(3*lam2^2*s2ahat^2 - 6*W1a^lam1*lam2^2*s2ahat^2 + 3*W1a^(2*lam1)*lam2^2*s2ahat^2) + s1ahat^2*(lam2^2*s2ahat^4 - 2*W1a^lam1*lam2^2*s2ahat^4 + W1a^(2*lam1)*lam2^2*s2ahat^4) - lam1*(cova^3*(2*lam2 + 2*lam2^2*m2ahat - 2*W1a^lam1*lam2 - 2*W2a^lam2*lam2 + 2*W1a^lam1*W2a^lam2*lam2 - 2*W1a^lam1*lam2^2*m2ahat) - cova^2*(6*lam2^2*m1ahat*s2ahat^2 - 6*W1a^lam1*lam2^2*m1ahat*s2ahat^2) + s1ahat^2*(cova*(6*lam2*s2ahat^2 + 6*lam2^2*m2ahat*s2ahat^2 - 6*W1a^lam1*lam2*s2ahat^2 - 6*W2a^lam2*lam2*s2ahat^2 - 6*W1a^lam1*lam2^2*m2ahat*s2ahat^2 + 6*W1a^lam1*W2a^lam2*lam2*s2ahat^2) - 2*lam2^2*m1ahat*s2ahat^4 + 2*W1a^lam1*lam2^2*m1ahat*s2ahat^4)))/(lam1^2*lam2^2*(s1ahat^2*s2ahat^2 - cova^2)^3)) h9_10=0 h9_11=sum(-((W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova^2*lam1 - 2*cova*lam2*s2ahat^2 + lam1*s1ahat^2*s2ahat^2 - W2a^lam2*cova^2*lam1 + 2*W1a^lam1*cova*lam2*s2ahat^2 + cova^2*lam1*lam2*m2ahat - W2a^lam2*lam1*s1ahat^2*s2ahat^2 + lam1*lam2*m2ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h9_12=sum(-((W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova^2*lam2 - 2*cova*lam1*s1ahat^2 + lam2*s1ahat^2*s2ahat^2 - W1a^lam1*cova^2*lam2 + 2*W2a^lam2*cova*lam1*s1ahat^2 + cova^2*lam1*lam2*m1ahat - W1a^lam1*lam2*s1ahat^2*s2ahat^2 + lam1*lam2*m1ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h10_1=0 h10_2=0 h10_3=sum((covb^2*(lam2*m2bhat - W2b^lam2 + 1) + s2bhat^2*((lam2*m2bhat - W2b^lam2 + 1)*s1bhat^2 - 2*covb*lam2*m1bhat))/(lam2*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s2bhat^2*(2*lam2 - 2*W1b^lam1*lam2))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_4=sum(- (covb*(s2bhat^4*(- 4*lam2^2*m1bhat^2*s1bhat + 2*lam2^2*s1bhat^3) - s1bhat^3*s2bhat^2*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2)) - covb^3*(s1bhat*(4*lam2*m2bhat - 4*W2b^lam2 + 2*W2b^(2*lam2) + 2*lam2^2*m2bhat^2 - 4*W2b^lam2*lam2*m2bhat + 2) + 2*lam2^2*s1bhat*s2bhat^2) + s1bhat^3*s2bhat^4*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + covb^2*s1bhat*s2bhat^2*(6*lam2*m1bhat + 6*lam2^2*m1bhat*m2bhat - 6*W2b^lam2*lam2*m1bhat))/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam1*((6*lam2 + 6*lam2^2*m2bhat - 6*W1b^lam1*lam2 - 6*W2b^lam2*lam2 + 6*W1b^lam1*W2b^lam2*lam2 - 6*W1b^lam1*lam2^2*m2bhat)*covb^2*s1bhat*s2bhat^2 + (8*W1b^lam1*lam2^2*m1bhat - 8*lam2^2*m1bhat)*covb*s1bhat*s2bhat^4 + (2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat)*s1bhat^3*s2bhat^4) - covb*s1bhat*s2bhat^4*(4*W1b^(2*lam1)*lam2^2 - 8*W1b^lam1*lam2^2 + 4*lam2^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_5=0 h10_6=0 h10_7=sum((covb^2*(lam1*m1bhat - W1b^lam1 + 1) + s1bhat^2*((lam1*m1bhat - W1b^lam1 + 1)*s2bhat^2 - 2*covb*lam1*m2bhat))/(lam1*(s1bhat^2*s2bhat^2 - covb^2)^2) - (covb*s1bhat^2*(2*lam1 - 2*W2b^lam2*lam1))/(lam1*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_8=sum(- (covb*(s1bhat^4*(- 4*lam1^2*m2bhat^2*s2bhat + 2*lam1^2*s2bhat^3) - s1bhat^2*s2bhat^3*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2)) - covb^3*(s2bhat*(4*lam1*m1bhat - 4*W1b^lam1 + 2*W1b^(2*lam1) + 2*lam1^2*m1bhat^2 - 4*W1b^lam1*lam1*m1bhat + 2) + 2*lam1^2*s1bhat^2*s2bhat) + s1bhat^4*s2bhat^3*(2*lam1*m2bhat + 2*lam1^2*m1bhat*m2bhat - 2*W1b^lam1*lam1*m2bhat) + covb^2*s1bhat^2*s2bhat*(6*lam1*m2bhat + 6*lam1^2*m1bhat*m2bhat - 6*W1b^lam1*lam1*m2bhat))/(lam1^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (lam2*((6*lam1 + 6*lam1^2*m1bhat - 6*W1b^lam1*lam1 - 6*W2b^lam2*lam1 + 6*W1b^lam1*W2b^lam2*lam1 - 6*W2b^lam2*lam1^2*m1bhat)*covb^2*s1bhat^2*s2bhat + (8*W2b^lam2*lam1^2*m2bhat - 8*lam1^2*m2bhat)*covb*s1bhat^4*s2bhat + (2*lam1 + 2*lam1^2*m1bhat - 2*W1b^lam1*lam1 - 2*W2b^lam2*lam1 + 2*W1b^lam1*W2b^lam2*lam1 - 2*W2b^lam2*lam1^2*m1bhat)*s1bhat^4*s2bhat^3) - covb*s1bhat^4*s2bhat*(4*W2b^(2*lam2)*lam1^2 - 8*W2b^lam2*lam1^2 + 4*lam1^2))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_9=0 h10_10=sum(- (s1bhat^2*(covb^2*(6*lam2*m2bhat - 6*W2b^lam2 + 3*W2b^(2*lam2) + 3*lam2^2*m2bhat^2 - 6*W2b^lam2*lam2*m2bhat + 3) - covb*(6*lam2*m1bhat*s2bhat^2 - 6*W2b^lam2*lam2*m1bhat*s2bhat^2 + 6*lam2^2*m1bhat*m2bhat*s2bhat^2) + lam2^2*m1bhat^2*s2bhat^4) - covb^3*(2*lam2*m1bhat + 2*lam2^2*m1bhat*m2bhat - 2*W2b^lam2*lam2*m1bhat) + s1bhat^4*(W2b^(2*lam2)*s2bhat^2 - 2*W2b^lam2*s2bhat^2 + s2bhat^2 - lam2^2*s2bhat^4 + 2*lam2*m2bhat*s2bhat^2 + lam2^2*m2bhat^2*s2bhat^2 - 2*W2b^lam2*lam2*m2bhat*s2bhat^2) + covb^4*lam2^2 + 3*covb^2*lam2^2*m1bhat^2*s2bhat^2)/(lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3) - (covb^2*(3*lam2^2*s2bhat^2 - 6*W1b^lam1*lam2^2*s2bhat^2 + 3*W1b^(2*lam1)*lam2^2*s2bhat^2) + s1bhat^2*(lam2^2*s2bhat^4 - 2*W1b^lam1*lam2^2*s2bhat^4 + W1b^(2*lam1)*lam2^2*s2bhat^4) - lam1*(covb^3*(2*lam2 + 2*lam2^2*m2bhat - 2*W1b^lam1*lam2 - 2*W2b^lam2*lam2 + 2*W1b^lam1*W2b^lam2*lam2 - 2*W1b^lam1*lam2^2*m2bhat) - covb^2*(6*lam2^2*m1bhat*s2bhat^2 - 6*W1b^lam1*lam2^2*m1bhat*s2bhat^2) + s1bhat^2*(covb*(6*lam2*s2bhat^2 + 6*lam2^2*m2bhat*s2bhat^2 - 6*W1b^lam1*lam2*s2bhat^2 - 6*W2b^lam2*lam2*s2bhat^2 - 6*W1b^lam1*lam2^2*m2bhat*s2bhat^2 + 6*W1b^lam1*W2b^lam2*lam2*s2bhat^2) - 2*lam2^2*m1bhat*s2bhat^4 + 2*W1b^lam1*lam2^2*m1bhat*s2bhat^4)))/(lam1^2*lam2^2*(s1bhat^2*s2bhat^2 - covb^2)^3)) h10_11=sum(-((W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb^2*lam1 - 2*covb*lam2*s2bhat^2 + lam1*s1bhat^2*s2bhat^2 - W2b^lam2*covb^2*lam1 + 2*W1b^lam1*covb*lam2*s2bhat^2 + covb^2*lam1*lam2*m2bhat - W2b^lam2*lam1*s1bhat^2*s2bhat^2 + lam1*lam2*m2bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h10_12=sum(-((W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb^2*lam2 - 2*covb*lam1*s1bhat^2 + lam2*s1bhat^2*s2bhat^2 - W1b^lam1*covb^2*lam2 + 2*W2b^lam2*covb*lam1*s1bhat^2 + covb^2*lam1*lam2*m1bhat - W1b^lam1*lam2*s1bhat^2*s2bhat^2 + lam1*lam2*m1bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_1=sum((2*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h11_2=sum((2*s1ahat*s2ahat^2*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_3=sum((2*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h11_4=sum((2*s1bhat*s2bhat^2*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_5=sum(-(cova*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1))/(lam1^2*(- cova^2 + s1ahat^2*s2ahat^2))) h11_6=sum(-(2*cova*s2ahat*(W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_7=sum(-(covb*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1))/(lam1^2*(- covb^2 + s1bhat^2*s2bhat^2))) h11_8=sum(-(2*covb*s2bhat*(W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_9=sum(-((W1a^lam1*lam1*log(W1a) - W1a^lam1 + 1)*(cova^2*lam1 - 2*cova*lam2*s2ahat^2 + lam1*s1ahat^2*s2ahat^2 - W2a^lam2*cova^2*lam1 + 2*W1a^lam1*cova*lam2*s2ahat^2 + cova^2*lam1*lam2*m2ahat - W2a^lam2*lam1*s1ahat^2*s2ahat^2 + lam1*lam2*m2ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m1ahat*s2ahat^2))/(lam1^3*lam2*(s1ahat^2*s2ahat^2 - cova^2)^2)) h11_10=sum(-((W1b^lam1*lam1*log(W1b) - W1b^lam1 + 1)*(covb^2*lam1 - 2*covb*lam2*s2bhat^2 + lam1*s1bhat^2*s2bhat^2 - W2b^lam2*covb^2*lam1 + 2*W1b^lam1*covb*lam2*s2bhat^2 + covb^2*lam1*lam2*m2bhat - W2b^lam2*lam1*s1bhat^2*s2bhat^2 + lam1*lam2*m2bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m1bhat*s2bhat^2))/(lam1^3*lam2*(s1bhat^2*s2bhat^2 - covb^2)^2)) h11_11=sum(((2*((W1a^lam1-1)/lam1^2-(W1a^lam1*log(W1a))/lam1)^2)/s1ahat^2-(2*(m1ahat-(W1a^lam1-1)/lam1)*((2*W1a^lam1-2)/lam1^3-(2*W1a^lam1*log(W1a))/lam1^2+(W1a^lam1*log(W1a)^2)/lam1))/s1ahat^2+(2*cova*(m2ahat-(W2a^lam2-1)/lam2)*((2*W1a^lam1-2)/lam1^3-(2*W1a^lam1*log(W1a))/lam1^2+(W1a^lam1*log(W1a)^2)/lam1))/(s1ahat^2*s2ahat^2))/((2*cova^2)/(s1ahat^2*s2ahat^2)-2))+sum(((2*((W1b^lam1-1)/lam1^2-(W1b^lam1*log(W1b))/lam1)^2)/s1bhat^2-(2*(m1bhat-(W1b^lam1-1)/lam1)*((2*W1b^lam1-2)/lam1^3-(2*W1b^lam1*log(W1b))/lam1^2+(W1b^lam1*log(W1b)^2)/lam1))/s1bhat^2+(2*covb*(m2bhat-(W2b^lam2-1)/lam2)*((2*W1b^lam1-2)/lam1^3-(2*W1b^lam1*log(W1b))/lam1^2+(W1b^lam1*log(W1b)^2)/lam1))/(s1bhat^2*s2bhat^2))/((2*covb^2)/(s1bhat^2*s2bhat^2)-2)) h11_12=sum((2*cova*(W1a^lam1*lam1*log(W1a)-W1a^lam1+1)*(W2a^lam2*lam2*log(W2a)-W2a^lam2+1))/(lam1^2*lam2^2*(-2*cova^2+2*s1ahat^2*s2ahat^2)))+sum((2*covb*(W1b^lam1*lam1*log(W1b)-W1b^lam1+1)*(W2b^lam2*lam2*log(W2b)-W2b^lam2+1))/(lam1^2*lam2^2*(-2*covb^2+2*s1bhat^2*s2bhat^2))) h12_1=sum(-(cova*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- cova^2 + s1ahat^2*s2ahat^2))) h12_2=sum(-(2*cova*s1ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam1 - lam2*s2ahat^2 - W2a^lam2*cova*lam1 + W1a^lam1*lam2*s2ahat^2 + cova*lam1*lam2*m2ahat - lam1*lam2*m1ahat*s2ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_3=sum(-(covb*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- covb^2 + s1bhat^2*s2bhat^2))) h12_4=sum(-(2*covb*s1bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam1 - lam2*s2bhat^2 - W2b^lam2*covb*lam1 + W1b^lam1*lam2*s2bhat^2 + covb*lam1*lam2*m2bhat - lam1*lam2*m1bhat*s2bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_5=sum((2*s1ahat^2*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1))/(lam2^2*(- 2*cova^2 + 2*s1ahat^2*s2ahat^2))) h12_6=sum((2*s1ahat^2*s2ahat*(W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova*lam2 - lam1*s1ahat^2 - W1a^lam1*cova*lam2 + W2a^lam2*lam1*s1ahat^2 + cova*lam1*lam2*m1ahat - lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_7=sum((2*s1bhat^2*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1))/(lam2^2*(- 2*covb^2 + 2*s1bhat^2*s2bhat^2))) h12_8=sum((2*s1bhat^2*s2bhat*(W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb*lam2 - lam1*s1bhat^2 - W1b^lam1*covb*lam2 + W2b^lam2*lam1*s1bhat^2 + covb*lam1*lam2*m1bhat - lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_9=sum(-((W2a^lam2*lam2*log(W2a) - W2a^lam2 + 1)*(cova^2*lam2 - 2*cova*lam1*s1ahat^2 + lam2*s1ahat^2*s2ahat^2 - W1a^lam1*cova^2*lam2 + 2*W2a^lam2*cova*lam1*s1ahat^2 + cova^2*lam1*lam2*m1ahat - W1a^lam1*lam2*s1ahat^2*s2ahat^2 + lam1*lam2*m1ahat*s1ahat^2*s2ahat^2 - 2*cova*lam1*lam2*m2ahat*s1ahat^2))/(lam1*lam2^3*(s1ahat^2*s2ahat^2 - cova^2)^2)) h12_10=sum(-((W2b^lam2*lam2*log(W2b) - W2b^lam2 + 1)*(covb^2*lam2 - 2*covb*lam1*s1bhat^2 + lam2*s1bhat^2*s2bhat^2 - W1b^lam1*covb^2*lam2 + 2*W2b^lam2*covb*lam1*s1bhat^2 + covb^2*lam1*lam2*m1bhat - W1b^lam1*lam2*s1bhat^2*s2bhat^2 + lam1*lam2*m1bhat*s1bhat^2*s2bhat^2 - 2*covb*lam1*lam2*m2bhat*s1bhat^2))/(lam1*lam2^3*(s1bhat^2*s2bhat^2 - covb^2)^2)) h12_11=sum((2*cova*(W1a^lam1*lam1*log(W1a)-W1a^lam1+1)*(W2a^lam2*lam2*log(W2a)-W2a^lam2+1))/(lam1^2*lam2^2*(-2*cova^2+2*s1ahat^2*s2ahat^2)))+sum((2*covb*(W1b^lam1*lam1*log(W1b)-W1b^lam1+1)*(W2b^lam2*lam2*log(W2b)-W2b^lam2+1))/(lam1^2*lam2^2*(-2*covb^2+2*s1bhat^2*s2bhat^2))) h12_12=sum(((2*((W2a^lam2-1)/lam2^2-(W2a^lam2*log(W2a))/lam2)^2)/s2ahat^2-(2*(m2ahat-(W2a^lam2-1)/lam2)*((2*W2a^lam2-2)/lam2^3-(2*W2a^lam2*log(W2a))/lam2^2+(W2a^lam2*log(W2a)^2)/lam2))/s2ahat^2+(2*cova*(m1ahat-(W1a^lam1-1)/lam1)*((2*W2a^lam2-2)/lam2^3-(2*W2a^lam2*log(W2a))/lam2^2+(W2a^lam2*log(W2a)^2)/lam2))/(s1ahat^2*s2ahat^2))/((2*cova^2)/(s1ahat^2*s2ahat^2)-2))+sum(((2*((W2b^lam2-1)/lam2^2-(W2b^lam2*log(W2b))/lam2)^2)/s2bhat^2-(2*(m2bhat-(W2b^lam2-1)/lam2)*((2*W2b^lam2-2)/lam2^3-(2*W2b^lam2*log(W2b))/lam2^2+(W2b^lam2*log(W2b)^2)/lam2))/s2bhat^2+(2*covb*(m1bhat-(W1b^lam1-1)/lam1)*((2*W2b^lam2-2)/lam2^3-(2*W2b^lam2*log(W2b))/lam2^2+(W2b^lam2*log(W2b)^2)/lam2))/(s1bhat^2*s2bhat^2))/((2*covb^2)/(s1bhat^2*s2bhat^2)-2)) h1r=c(h1_1,h1_2,h1_3,h1_4,h1_5,h1_6,h1_7,h1_8,h1_9,h1_10,h1_11,h1_12) h2r=c(h2_1,h2_2,h2_3,h2_4,h2_5,h2_6,h2_7,h2_8,h2_9,h2_10,h2_11,h2_12) h3r=c(h3_1,h3_2,h3_3,h3_4,h3_5,h3_6,h3_7,h3_8,h3_9,h3_10,h3_11,h3_12) h4r=c(h4_1,h4_2,h4_3,h4_4,h4_5,h4_6,h4_7,h4_8,h4_9,h4_10,h4_11,h4_12) h5r=c(h5_1,h5_2,h5_3,h5_4,h5_5,h5_6,h5_7,h5_8,h5_9,h5_10,h5_11,h5_12) h6r=c(h6_1,h6_2,h6_3,h6_4,h6_5,h6_6,h6_7,h6_8,h6_9,h6_10,h6_11,h6_12) h7r=c(h7_1,h7_2,h7_3,h7_4,h7_5,h7_6,h7_7,h7_8,h7_9,h7_10,h7_11,h7_12) h8r=c(h8_1,h8_2,h8_3,h8_4,h8_5,h8_6,h8_7,h8_8,h8_9,h8_10,h8_11,h8_12) h9r=c(h9_1,h9_2,h9_3,h9_4,h9_5,h9_6,h9_7,h9_8,h9_9,h9_10,h9_11,h9_12) h10r=c(h10_1,h10_2,h10_3,h10_4,h10_5,h10_6,h10_7,h10_8,h10_9,h10_10,h10_11,h10_12) h11r=c(h11_1,h11_2,h11_3,h11_4,h11_5,h11_6,h11_7,h11_8,h11_9,h11_10,h11_11,h11_12) h12r=c(h12_1,h12_2,h12_3,h12_4,h12_5,h12_6,h12_7,h12_8,h12_9,h12_10,h12_11,h12_12) HCF=rbind(h1r,h2r,h3r,h4r,h5r,h6r,h7r,h8r,h9r,h10r,h11r,h12r) HCF[1,2]=0;HCF[1,6]=0;HCF[1,9]=0; HCF[2,1]=0;HCF[2,5]=0; HCF[3,4]=0;HCF[3,8]=0;HCF[3,10]=0;HCF[4,3]=0; HCF[4,7]=0; HCF[5,2]=0;HCF[5,6]=0;HCF[5,9]=0; HCF[6,1]=0;HCF[6,5]=0; HCF[7,4]=0;HCF[7,8]=0;HCF[7,10]=0; HCF[8,3]=0;HCF[8,7]=0; HCF[9,1]=0;HCF[9,5]=0; HCF[10,3]=0; HCF[10,7]=0; HCF[1,1]=na*HCF[1,1]; HCF[1,5]=na*HCF[1,5]; HCF[5,1]=na*HCF[5,1]; HCF[3,3]=nb*HCF[3,3]; HCF[3,7]=nb*HCF[3,7]; HCF[7,3]=nb*HCF[7,3]; HCF[7,7]=nb*HCF[7,7]; HCF[5,5]=na*HCF[5,5]; VV=inv(-HCF) att=1-atSpec; t=att; s1alam=sqrt( var(W1alam)*(na-1)/na ) s1blam=sqrt( var(W1blam)*(nb-1)/nb ) s2alam=sqrt( var(W2alam)*(na-1)/na ) s2blam=sqrt( var(W2blam)*(nb-1)/nb ) m1alam=mean(W1alam) m1blam=mean(W1blam) m2alam=mean(W2alam) m2blam=mean(W2blam) ############################################################################# ############################################################################# #==================AT A GIVEN T FOR THE ORIGINAL=================== ROC1or= (1-pnorm(qnorm(1-t,m1alam,s1alam),m1blam,s1blam)); ROC2or= (1-pnorm(qnorm(1-t,m2alam,s2alam),m2blam,s2blam)); dm1a = -(2^(1/2)*exp(-((m1ahat - m1bhat)/s1bhat + (2^(1/2)*s1ahat*erfcinv(2*t))/s1bhat)^2/2))/(2*s1bhat*pi^(1/2)) dm1b = (2^(1/2)*exp(-((m1ahat - m1bhat)/s1bhat + (2^(1/2)*s1ahat*erfcinv(2*t))/s1bhat)^2/2))/(2*s1bhat*pi^(1/2)) ds1a = -(erfcinv(2*t)*exp(-((m1ahat - m1bhat)/s1bhat + (2^(1/2)*s1ahat*erfcinv(2*t))/s1bhat)^2/2))/(s1bhat*pi^(1/2)) ds1b = (2^(1/2)*exp(-((m1ahat - m1bhat)/s1bhat + (2^(1/2)*s1ahat*erfcinv(2*t))/s1bhat)^2/2)*((m1ahat - m1bhat)/s1bhat^2 + (2^(1/2)*s1ahat*erfcinv(2*t))/s1bhat^2))/(2*pi^(1/2)) der1=c(dm1a, ds1a, dm1b, ds1b); vROC1=c(der1)%*%VV[1:4,1:4]%*%t(t(c(der1))) vROC1=c(vROC1) dm2a = -(2^(1/2)*exp(-((m2ahat - m2bhat)/s2bhat + (2^(1/2)*s2ahat*erfcinv(2*t))/s2bhat)^2/2))/(2*s2bhat*pi^(1/2)) dm2b = (2^(1/2)*exp(-((m2ahat - m2bhat)/s2bhat + (2^(1/2)*s2ahat*erfcinv(2*t))/s2bhat)^2/2))/(2*s2bhat*pi^(1/2)) ds2a = -(erfcinv(2*t)*exp(-((m2ahat - m2bhat)/s2bhat + (2^(1/2)*s2ahat*erfcinv(2*t))/s2bhat)^2/2))/(s2bhat*pi^(1/2)) ds2b = (2^(1/2)*exp(-((m2ahat - m2bhat)/s2bhat + (2^(1/2)*s2ahat*erfcinv(2*t))/s2bhat)^2/2)*((m2ahat - m2bhat)/s2bhat^2 + (2^(1/2)*s2ahat*erfcinv(2*t))/s2bhat^2))/(2*pi^(1/2)) der2=c(dm2a, ds2a, dm2b, ds2b); vROC2=c(der2)%*%VV[5:8,5:8]%*%t(t(c(der2))) covROC=c(der1, 0, 0, 0, 0)%*%VV[1:8,1:8]%*%t(t(c(0, 0, 0, 0, der2))) Z=(ROC2or-ROC1or)/(sqrt(vROC2+vROC1-2*covROC)) SEZ=sqrt(vROC1+vROC2-2*covROC) pval2tZ=2*pnorm(-abs(Z)); CIoriginal=c((ROC2or-ROC1or)-pmalpha*SEZ, (ROC2or-ROC1or)+pmalpha*SEZ) CIoriginal ############################################################################# ############################################################################# #==================AT A GIVEN T FOR THE TRANSORMED=================== ROC1MC= qnorm(1-pnorm(qnorm(1-t,m1alam,s1alam),m1blam,s1blam)); ROC2MC= qnorm(1-pnorm(qnorm(1-t,m2alam,s2alam),m2blam,s2blam)); dm1a =-1/s1blam; ds1a =-(2^(1/2)*erfcinv(2*t))/s1blam; dm1b =1/s1blam; ds1b =(m1alam - m1blam)/s1blam^2 + (2^(1/2)*s1alam*erfcinv(2*t))/s1blam^2; der1=c(dm1a, ds1a, dm1b, ds1b); vROC1=c(der1)%*%VV[1:4,1:4]%*%t(t(c(der1))) vROC1=c(vROC1) dm2a =-1/s2blam; ds2a =-(2^(1/2)*erfcinv(2*t))/s2blam; dm2b =1/s2blam; ds2b =(m2alam - m2blam)/s2blam^2 + (2^(1/2)*s2alam*erfcinv(2*t))/s2blam^2; der2=c(dm2a, ds2a, dm2b, ds2b); vROC2=c(der2)%*%VV[5:8,5:8]%*%t(t(c(der2))) vROC2=c(vROC2) covROC=c(der1, 0, 0, 0, 0)%*%VV[1:8,1:8]%*%t(t(c(0, 0, 0, 0, der2))) covROC=c(covROC) Zroct=(ROC2MC-ROC1MC)/(sqrt(vROC2+vROC1-2*covROC)) SEroct=sqrt(vROC1+vROC2-2*covROC) pval2t_roct=2*pnorm(-abs(Zroct));pval2t=pval2t_roct; CIZstar=c((ROC2MC-ROC1MC)-pmalpha*SEroct, (ROC2MC-ROC1MC)+pmalpha*SEroct) CIZstar SE1=pnorm(ROC1MC) SE2=pnorm(ROC2MC) res <- data.frame(Sens1 = formattable(SE1, digits = 4, format = "f"), Sens2 = formattable(SE2, digits = 4, format = "f"), diff = formattable(SE2 - SE1, digits = 4, format = "f"), p_value_probit = formattable(pval2t, digits = 4, format = "f"), p_value = formattable(pval2tZ, digits = 4, format = "f"), ci_ll = formattable(CIoriginal[1], digits = 4, format = "f"), ci_ul = formattable(CIoriginal[2], digits = 4, format = "f")) rownames(res) <- c("Estimates:") colnames(res) <- c("Sens 1", "Sens 2", "Diff", "P-Val (Probit)", "P-Val", "CI (LL)", "CI (UL)") res <- formattable(as.matrix(res), digits = 4, format = "f") res #====TWO ROC FUNCTIONS: ONE IS THE BOXCOX AND THE OTHER THE REFERENCE LINE================ roc1<-function(t){ na = length(W1alam) nb = length(W1blam) s1alam=sqrt( var(W1alam)*(na-1)/na ) s1blam=sqrt( var(W1blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W1alam), sd=s1alam), mean=mean(W1blam), sd=s1blam) } roc2<-function(t){ na = length(W2alam) nb = length(W2blam) s2alam=sqrt( var(W2alam)*(na-1)/na ) s2blam=sqrt( var(W2blam)*(nb-1)/nb ) 1-pnorm(qnorm(1-t, mean=mean(W2alam), sd=s2alam), mean=mean(W2blam), sd=s2blam) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } #================IF PLOTS ARE REQUESTED PLOT THE ROCS== if (plots=="on") { # x11() plot(linspace(0,1,1000),roc1(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="red") lines(linspace(0,1,1000),roc2(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="black") lines(linspace(0,1,1000),linspace(0,1,1000),type="l", lty=2) points(1-atSpec,SE1, col = "red") points(1-atSpec,SE2, col = "black") legend("bottomright", legend=c(paste("ROC for Marker 1 with Sens =", formattable(SE1, digits = 4, format = "f"), "at Spec =", formattable(atSpec, digits = 4, format = "f")), paste("ROC for Marker 2 with Sens =", formattable(SE2, digits = 4, format = "f"), "at Spec =", formattable(atSpec, digits = 4, format = "f")), paste("P-value for the Sens difference:", formattable(pval2t, digits = 4, format = "f"))), col=c("red", "black", "white"), lty=c(1, 1, NA), pch = c(NA, NA, NA), cex=0.8) } return(list(resultstable=res, Sens1=SE1, Sens2=SE2, pvalue_probit_difference= pval2t, pvalue_difference= pval2tZ, CI_difference= CIoriginal, roc1=roc1, roc2=roc2, transx1=W1alam, transy1=W1blam, transformation.parameter.1 = lam[1], transx2=W2alam, transy2=W2blam, transformation.parameter.2 = lam[2])) } } ``` ```{r, echo=FALSE, eval=TRUE} comparebcSpec <-function (marker1, marker2, D, atSens, alpha, plots){ if ((length(marker1) != length(D)) | (length(marker2) != length(D))) { stop("ERROR: The length of the 'marker' and 'D' inputs must be equal.") } else if (min(D) != 0 | max(D) != 1) { stop("ERROR: Controls must be assigned a value of 0; cases must be assigned a value of 1. Both controls and cases should be included in the dataset.") } else if (alpha <= 0 | alpha >= 1) { stop("ERROR: The level of significance, alpha, should be set between 0 and 1. A common choice is 0.05.") } else if (sum(is.na(marker1)) > 0 | sum(is.na(marker2)) > 0 | sum(is.na(D)) > 0 | sum(is.na(atSens)) > 0) { stop("ERROR: Please remove all missing data before running this function.") } else if ((!is.numeric(atSens)) | (length(atSens) != 1)) { stop("ERROR: 'atSens' must be a single numeric value.") } else if ((sum(marker1 < 0) > 0) | (sum(marker2 < 2)) > 0) { stop("ERROR: To use the Box-Cox transformation, all marker values must be positive.") } else { erf <- function (x) 2 * pnorm(x * sqrt(2)) - 1 erfc <- function (x) 2 * pnorm(x * sqrt(2), lower.tail = FALSE) erfinv <- function (x) qnorm((1 + x)/2)/sqrt(2) erfcinv <- function (x) qnorm(x/2, lower.tail = FALSE)/sqrt(2) pmalpha=qnorm(1-alpha/2); if (plots!="on"){plots="off"} W1a=marker1[D==0] W1b=marker1[D==1] W2a=marker2[D==0] W2b=marker2[D==1] Dflip=abs(D-1); atSpec=1-atSens; #out=comparebcSens(marker1group1=W1b, marker1group2=W1a ,marker2group1=W2b, marker2group2=W2a, atSpec=atSens, plots="off") out=comparebcSens(marker1, marker2, Dflip, atSpec=1-atSens, alpha, plots="off") #return(list(resultstable=res,Sens1=SE1,Sens2=SE2, pvalue_probit_difference= pval2t, CI_probit_difference= CIZstar, pvalue_difference= pval2tZ, CI_difference= CIoriginal, roc1=roc1, roc2=roc2, transx1=W1alam, transy1=W1blam, transx2=W2alam, transy2=W2blam)) FPR1=out$Sens1 FPR2=out$Sens2 CIZstar=out$CI_probit_difference CIoriginal=out$CI_difference pval2t=out$pvalue_probit_difference pval2tZ=out$pvalue_difference W1alam=out$transx1 W1blam=out$transy1 W2alam=out$transx2 W2blam=out$transy2 #====TWO ROC FUNCTIONS: ONE IS THE BOXCOX AND THE OTHER THE REFERENCE LINE================ roct1<-function(t){ na = length(W1alam) nb = length(W1blam) s1alam=sqrt(var(W1alam)*(na-1)/na) s1blam=sqrt(var(W1blam)*(nb-1)/nb) 1-pnorm(qnorm(1-t, mean=mean(W1blam), sd=s1blam), mean=mean(W1alam), sd=s1alam) } roct2<-function(t){ na = length(W2alam) nb = length(W2blam) s2alam=sqrt(var(W2alam)*(na-1)/na) s2blam=sqrt(var(W2blam)*(nb-1)/nb) 1-pnorm(qnorm(1-t, mean=mean(W2blam), sd=s2blam), mean=mean(W2alam), sd=s2alam) } rocuseless<-function(t){ 1-pnorm(qnorm(1-t,mean=1,sd=1),mean=1,sd=1) } #================IF PLOTS ARE REQUESTED PLOT THE ROCS== if (plots=="on") { # x11() plot(linspace(0,1,1000),roct1(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="red") lines(linspace(0,1,1000),roct2(linspace(0,1,1000)),main="Box-Cox Based ROCs",xlab="FPR = 1 - Specificity",ylab="TPR = Sensitivity",type="l",col="black") lines(linspace(0,1,1000),linspace(0,1,1000),type="l", lty=2) points(FPR1, 1-atSpec, col = "red") points(FPR2, 1-atSpec, col = "black") #line for the Youden: #lines(c(1-atSpec,1-SE1),c(1-atSpec,1-SE2),col="green") legend("bottomright", legend=c(paste("ROC for Marker 1 with FPR =", formattable(FPR1, digits = 4, format = "f"), "at Sens =", formattable(atSens, digits = 4, format = "f")), paste("ROC for Marker 2 with FPR =", formattable(FPR2, digits = 4, format = "f"), "at Sens =", formattable(atSens, digits = 4, format = "f")), paste("P-value for the Spec difference:", formattable(pval2t, digits = 4, format = "f"))), col=c("red", "black", "white"), lty=c(1, 1, NA), pch = c(NA, NA, NA), cex=0.8) } res <- data.frame(FPR1 = formattable(FPR1, digits = 4, format = "f"), FPR2 = formattable(FPR2, digits = 4, format = "f"), formattable(FPR2 - FPR1, digits = 4, format = "f"), p_value_probit = formattable(pval2t, digits = 4, format = "f"), p_value = formattable(pval2tZ, digits = 4, format = "f"), ci_ll = formattable(CIoriginal[1], digits = 4, format = "f"), ci_ul = formattable(CIoriginal[2], digits = 4, format = "f")) rownames(res) <- c("Estimates:") colnames(res) <- c("FPR 1", "FPR 2", "Diff", "P-Val (Probit)", "P-Val", "CI (LL)", "CI (UL)") res <- formattable(as.matrix(res), digits = 4, format = "f") res #return(list(resultstable=res,Sens1=SE1,Sens2=SE2, pvalue_probit_difference= pval2t, CI_probit_difference= CIZstar, pvalue_difference= pval2tZ, CI_difference= CIoriginal, roc1=roc1, roc2=roc2, transx1=W1alam, transy1=W1blam, transx2=W2alam, transy2=W2blam)) return(list(resultstable=res, FPR1=FPR1, FPR2=FPR2, pvalue_probit_difference= pval2t, pvalue_difference= pval2tZ, CI_difference= CIoriginal, roc1=roct1, roc2=roct2, transx1=W1alam, transy1=W1blam, transformation.parameter.1 = out$transformation.parameter.1, transx2=W2alam, transy2=W2blam, transformation.parameter.2 = out$transformation.parameter.2)) #return(list(FPR1=FPR1)) } } ``` ```{r, echo=FALSE, eval=TRUE} threerocs2 <- function (marker1, marker2, D, plots) { erf <- function (x) 2 * pnorm(x * sqrt(2)) - 1 erfc <- function (x) 2 * pnorm(x * sqrt(2), lower = FALSE) erfinv <- function (x) qnorm((1 + x)/2)/sqrt(2) erfcinv <- function (x) qnorm(x/2, lower = FALSE)/sqrt(2) W1a=marker1[D==0] W1b=marker1[D==1] W2a=marker2[D==0] W2b=marker2[D==1] Wa=cbind(W1a,W2a); Wb=cbind(W1b,W2b); na=length(W1a) nb=length(W1b) likbox2D<-function(W1a,W1b,W2a,W2b,lam){ na=length(W1a); nb=length(W1b); out=c(); #for (i in 1:length(h)){ W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) cova= cov(Walam)*(na-1)/na; covb= cov(Wblam)*(nb-1)/nb; mualam=c(mean(W1alam), mean(W2alam)) mublam=c(mean(W1blam), mean(W2blam)) out= -sum(log(dmvnorm(Walam,mualam,cova)))-sum(log(dmvnorm(Wblam,mublam,covb))) -(lam[1]-1)*sum(log(W1a))-(lam[2]-1)*sum(log(W2a))-(lam[1]-1)*sum(log(W1b))-(lam[2]-1)*sum(log(W2b)); return(out) } lam=c(2,2) likbox2D(W1a,W1b,W2a,W2b,lam) logL<-function(h){ likbox2D(W1a,W1b,W2a,W2b,h) } init=c(0.8,0.8) lam=fminsearch(logL,init) lam=c(lam$optbase$xopt) lam out <- optim(c(1,1), logL, method = "Nelder-Mead") lam = out$par ################################### W1alam= (W1a^lam[1]-1)/lam[1]; W2alam= (W2a^lam[2]-1)/lam[2]; W1blam= (W1b^lam[1]-1)/lam[1]; W2blam= (W2b^lam[2]-1)/lam[2]; Walam=cbind(W1alam,W2alam) Wblam=cbind(W1blam,W2blam) lam1=lam[1] lam2=lam[2] cova= cov(Walam)*(na-1)/na;cova=cova[1,2]; covb= cov(Wblam)*(nb-1)/nb;covb=covb[1,2]; s1alam=sqrt( var(W1alam)*(na-1)/na ) s1blam=sqrt( var(W1blam)*(nb-1)/nb ) s2alam=sqrt( var(W2alam)*(na-1)/na ) s2blam=sqrt( var(W2blam)*(nb-1)/nb ) m1alam=mean(W1alam) m1blam=mean(W1blam) m2alam=mean(W2alam) m2blam=mean(W2blam) roc1bc<-function(t){ 1-pnorm(qnorm(1-t,mean=mean(W1alam),sd=s1alam),mean=mean(W1blam),sd=s1blam) } roc2bc<-function(t){ 1-pnorm(qnorm(1-t,mean=mean(W2alam),sd=s2alam),mean=mean(W2blam),sd=s2blam) } ####################### MRM ################### rating=c(marker1,marker2) truth=c(D,D) reader=c(linspace(1,1,length(marker1)),linspace(2,2,length(marker2))) curves_binorm <- roc_curves(truth, rating, groups = list(Reader = reader), method = "binormal") params_binorm <- parameters(curves_binorm) t=linspace(0,1,1000); roct_mrm_1=pnorm(c(params_binorm$a[1])+c(params_binorm$b[1])*qnorm(t)) roct_mrm_2=pnorm(c(params_binorm$a[2])+c(params_binorm$b[2])*qnorm(t)) # Plot the empirical ROC curve roc_obj1 <- roc(D, marker1, quiet = TRUE) auc_emp1 <- auc(roc_obj1) tpr1=roc_obj1$sensitivities fpr1=1-roc_obj1$specificities roc_obj2 <- roc(D, marker2, quiet = TRUE) auc_emp2 <- auc(roc_obj2) tpr2=roc_obj2$sensitivities fpr2=1-roc_obj2$specificities AUC_empirical_marker1=auc_emp1 AUC_empirical_marker2=auc_emp2 if (plots == "on") { plot(t, roc1bc(t), type="l", lty = "dotted", col="indianred2", lwd=1.5, ylab="TPR = Sensitivity", xlab="FPR = 1 - Sensitivity") lines(t, roc2bc(t), type="l", lty = "dotted", col="red4", lwd=1.5) lines(t, roct_mrm_1, type="l", lty = "dashed", lwd=1.5, col="green2") lines(t, roct_mrm_2, type="l", lty = "dashed", lwd=1.5, col="darkgreen") lines(fpr1, tpr1, type="l", lty = "solid", lwd=1, col="deepskyblue") lines(fpr2, tpr2, type="l",lty = "solid", lwd=1, col="blue4") } #plot(fpr1, tpr1, type = "l", col = "magenta", lwd = 2, # xlim = c(0, 1), ylim = c(0, 1), xlab = "False Positive Rate (FPR)", # ylab = "True Positive Rate (TPR)", main = "Three Estimates of the ROC Curve") AUC1bc=trapz(t,roc1bc(t)) AUC2bc=trapz(t,roc2bc(t)) AUCmetz1=trapz(t,roct_mrm_1) AUCmetz2=trapz(t,roct_mrm_2) AUCemp1=abs(trapz(fpr1,tpr1)) AUCemp2=abs(trapz(fpr2,tpr2)) #legend("bottomright", legend = c("Box-Cox marker 1", "Box-Cox marker 2", "Metz marker 1",, "Metz marker 2","Empirical marker 1","Empirical marker 2"), # col = c("black","red","blue","green", "magenta", "cyan"), lty = c("solid", "solid", "dashed","dashed","dashed","dashed"), # lwd = 2, bty = "n") #CORRECT ONE: if (plots == "on") { legend("bottomright", legend = c(paste("Box-Cox Marker 1 (AUC = ", formattable(AUC1bc, digits = 4, format = "f"), ")", sep = ""), paste("Box-Cox Marker 2 (AUC = ", formattable(AUC2bc, digits = 4, format = "f"), ")", sep = ""), paste("Metz Marker 1 (AUC = ", formattable(AUCmetz1, digits = 4, format = "f"), ")", sep = ""), paste("Metz Marker 2 (AUC = ", formattable(AUCmetz2, digits = 4, format = "f"), ")", sep = ""), paste("Empirical Marker 1 (AUC = ", formattable(AUCemp1, digits = 4, format = "f"), ")", sep = ""), paste("Empirical Marker 2 (AUC = ", formattable(AUCemp2, digits = 4, format = "f"), ")", sep = "")), col = c("indianred2", "red4", "green2", "darkgreen", "deepskyblue", "blue4"), lty = c("dotted", "dotted", "dashed", "dashed", "solid", "solid"), cex = 0.8, lwd = 2) title(main = "Three Estimates of the ROC Curve (Two Markers)") } return(list(AUC_BoxCox1 = AUC1bc, AUC_BoxCox2 = AUC2bc, AUC_Metz1 = AUCmetz1, AUC_Metz2 = AUCmetz2, AUC_Empirical1 = AUCemp1, AUC_Empirical2 = AUCemp2)) } ``` ## About the rocbc Package (Authors: LE Bantis, B Brewer, CT Nakas, B Reiser) This package is focused on Box-Cox based ROC curves and provides point estimates, confidence intervals (CIs), and hypothesis tests. It can be used both for inferences for a single biomarker and when comparisons of two correlated biomarkers are of interest. It provides inferences and comparisons around the area under the ROC curve (AUC), the Youden index, the sensitivity at a given specificity level (and vice versa), the optimal operating point of the ROC curve (in the Youden sense), and the Youden based cutoff. This documentation consists of two parts, the one marker and the two marker case. All approaches presented herein have been recently published (see references for each function). The functions of each part: * **One marker case** + checkboxcox + threerocs + rocboxcox + rocboxcoxCI * **Two marker case (comparisons of two markers)** + checkboxcox2 + comparebcAUC + comparebcJ + comparebcSens + comparebcSpec + threerocs2 --- **checkboxcox** *** * **Description** + This function tests whether the Box-Cox transformation is able to achieve approximate normality for your data. That is, it will allow the user to know whether it is appropriate to use all the methods discussed later on in this package. * **Usage** ```checkboxcox(marker, D, plots, printShapiro = TRUE) ``` * **Arguments** + `marker`: A vector of length $n$ that contains the biomarker scores of all individuals. + ```D```: A vector of length n that contains the true disease status of an individual. It is a binary vector containing 0 for the healthy/control individuals and 1 for the diseased individuals. + ```plots='on' or 'off'```: Valid inputs are "on" and "off". When set to "on", the user gets the histograms of the biomarker for both the healthy and the diseased group before and after the Box-Cox transformation. In addition, all four corresponding qq-plots are provided. + ```printShapiro```: Boolean. When set to TRUE, the results of the Shapiro-Wilk test will be printed to the console. When set to FALSE, the results are suppressed. Default value is FALSE. * **Value** + ```res_shapiro```: A results table that contains the results of four Shapiro-Wilk tests for normality testing. Two of these refer to normality testing of the healthy and the diseased groups before the Box-Cox transformation, and the remaining two refer to the Box-Cox transformed biomarkers scores for the healthy and the diseased groups. Thus, this testing process produces four p-values. In addition, if the plots are set to 'on', then the output provides (1) the histograms of the biomarker for both the healthy and the diseased groups before and after the Box-Cox transformation, (2) all four corresponding qq-plots, and (3) a plot with the empirical ROC curve overplotted with the Box-Cox based ROC curve for visual comparison purposes. + ```transformation.parameter```: The single transformation parameter $\lambda$ that is applied for both groups simultaneously. + ```transx```: The Box-Cox transformed scores for the healthy. + ```transy```: The Box-Cox transformed scores for the diseased. + ```pval_x```: The p-value of the Shapiro-Wilk test of normality for the healthy group (before the Box-Cox transformation). + ```pval_y```: The p-value of the Shapiro-Wilk test of normality for the diseased group (before the Box-Cox transformation). + ```pval_transx```: The p-value of the Shapiro-Wilk test of normality for the healthy group (after the Box-Cox transformation). + ```pval_transy```: The p-value of the Shapiro-Wilk test of normality for the diseased group (after the Box-Cox transformation). + ```roc```: A function of the estimated Box-Cox ROC curve. You can use this to simply request TPR values for given FPR values. * **Example: ** ```{r, echo=TRUE, eval=TRUE} #DATA GENERATION set.seed(123) x=rgamma(100, shape=2, rate = 8) # Generates biomarker data from a gamma distribution # for the healthy group. y=rgamma(100, shape=2, rate = 4) # Generates biomarker data from a gamma distribution # for the diseased group. scores=c(x,y) D=c(zeros(1,100), ones(1,100)) out=checkboxcox(marker=scores, D, plots="on", printShapiro = TRUE) summary(out) ``` --- **threerocs** *** * **Description** + This function provides a visual comparison of the Empirical ROC, the Box-Cox ROC, and the Metz binormal semi-parametric estimator of the ROC curve. It also computes the AUC for the curve corresponding to each method. * **Usage** ```threerocs(marker, D, plots) ``` * **Arguments** + `marker`: A vector of length $n$ that contains the biomarker scores of all individuals. + ```D```: A vector of length n that contains the true disease status of an individual. It is a binary vector containing 0 for the healthy/control individuals and 1 for the diseased individuals. + ```plots='on' or 'off'```: Valid inputs are "on" and "off". When set to "on", the user gets a single plot containing the estimated ROC curves using the Empirical, Box-Cox, and Metz methods. * **Value** + ```AUC_Empirical```: The AUC of the empirical ROC curve. + ```AUC_Metz```: The AUC of the Metz binormal curve (as calculated by MRMCaov package using the "binormal" option). + ```AUC_BoxCox```: The AUC of the Box-Cox based ROC curve. * **Example: ** ```{r, echo=TRUE, eval=TRUE} #DATA GENERATION set.seed(123) x <- rgamma(100, shape=2, rate = 8) # generates biomarker data from a gamma # distribution for the healthy group. y <- rgamma(100, shape=2, rate = 4) # generates biomarker data from a gamma # distribution for the diseased group. scores <- c(x,y) D=c(pracma::zeros(1,100), pracma::ones(1,100)) out=threerocs(marker=scores, D, plots="on") summary(out) ``` --- **rocboxcox** *** * **Description** + This function applies the Box-Cox transformation to provide a comprehensive ROC analysis that involves the AUC (and its CI), the maximized Youden index (and its CI), the optimized Youden based cutoff (and its CI), and joint confidence regions for the optimized pair of sensitivity and specificity. For the AUC and the Youden index, the Delta Method is employed by accounting for the variability of the estimated transformation parameters due to the Box-Cox transformation. Both the AUC and the YI confidence intervals are found after using the probit transformation and back-transforming the endpoints of the corresponding CI back into the ROC space. For the cutoffs, it has been shown that the delta method does not perform well; thus, the bootstrap with 1,000 repetitions is employed here, instead. During this this process, the cutoffs are back-transformed with the inverse Box-Cox transformation so that they lie on the original scale of the data. * **Usage** ```rocboxcox(marker, D, plots, printProgress = TRUE) ``` * **Arguments** + `marker`: A vector of length n that contains the biomarker scores of all individuals. + ```D```: A vector of length n that contains the true disease status of an individual. It is a binary vector containing 0 for the healthy/control individuals and 1 for the diseased individuals. + ```alpha```: Nominal level used to calculate the confidence intervals. A common choice is 0.05. + ```plots```: Valid inputs are "on" and "off". When set to "on", the function returns a comprehensive figure with the ROC estimate and several point estimates: AUC, Youden index, optimized Youden cutoff, and the Youden based sensitivty and specificity along with the corresponding marginal confidence intervals and the joint confidence region of the estimated sensitivity and specificity. + ```printProgress```: Boolean. When set to TRUE, messages describing the progress of the bootstrapping will be printed to the console window. When set to FALSE, these messages are suppressed. Default value is FALSE. * **Value** + ```transx```: The Box-Cox transformed scores for the healthy group. + ```transy```: The Box-Cox transformed scores for the diseased group. + ```transformation.parameter```: The estimated Box-Cox transformation parameter ($\lambda$). + ```AUC```: The estimated area under the Box-Cox based ROC curve. + ```AUCCI```: The (1-$\alpha$)100\% CI for the AUC. (A common choice of $\alpha$ is 0.05). This CI is based on probit transforming the AUC estimate, finding the CI on the real line, and then back-transforming its endpoints to the ROC space. It is in line with using $Z=\frac{\Phi^{-1}(\hat{AUC})}{\sqrt{var(\Phi^{-1}(\hat{AUC}))}}$ to test the null hypothesis that $H_{0}: AUC=0.5$. + ```pvalueAUC```: The corresponding p-value for the AUC estimate. This a two-tailed p-value that tests the hypothesis $H_{0}: AUC=0.5$ by employing $Z=\frac{\Phi^{-1}(\hat{AUC})}{\sqrt{var(\Phi^{-1}(\hat{AUC}))}}$ which, under the null hypothesis, is taken to follow a standard normal distribution. + ```J```: The maximized Youden index. + ```JCI```: The corresponding CI for the maximized Youden index. For this CI, we consider the probit transformation $\hat{J}_{T}=\Phi^{-1}(\hat{J})$ and then back-transform its endpoints to derive a 95\% CI for the Youden index itself (Bantis et al. (2021)). + ```pvalueJ```: The corresponding two-tailed p-value for the maximized Youden index. This corresponds to $Z=\frac{\hat{J}_{T}}{\sqrt{var(\hat{J}_{T})}}$. The underlying null hypothesis is $H_{0}: J=0$ + ```Sens```: The sensitivity that corresponds to the Youden based optimized cutoff. + ```CImarginalSens```: The marginal (1-$\alpha$)100\% CI for the sensitivity that corresponds to the Youden based optimized cutoff. This is derived by first employing the probit transformation, finding a CI on the real line, and then back-transforming its endpoints to the ROC space. + ```Spec```: The specificity that corresponds to the Youden based optimized cutoff. + ```CImarginalSpec```: The marginal (1-$\alpha$)100\% CI for the specificity that corresponds to the Youden based optimized cutoff. This is derived by first employing the probit transformation, finding a CI on the real line, and then back-transforming its endpoints to the ROC space. + ```cutoff```: The Youden-based optimized cutoff. + ```CIcutoff```: the (1-$\alpha$)100\% CI for the Youden-based optimized cutoff. This is based on the bootstrap. It involves using the Box-Cox transformation for every bootstrap iteration and then using the inverse Box-Cox transformation to obtain the cutoff on its original scale (Bantis et al. (2019)). + ```areaegg```: The area of the (1-$\alpha$)100\% egg-shaped joint confidence region that refers to the optimized pair of sensitivity and specificity. This takes into account the fact that the estimated sensitivity and specificity are correlated as opposed to the corresponding rectangular area that ignores this. + ```arearect```: The area of the (1-$\alpha$)100\% rectangular joint confidence region that refers to the optimized pair of sensitivity and specificity. This ignores the correlation of the optimized sensitivity and specificity and tends to yield a larger area compared to the one of the egg-shaped region. + ```mxlam```: The mean of the marker scores of the healthy group after the Box-Cox transformation. + ```sxlam```: The standard deviation of the marker scores of the healthy group after the Box-Cox transformation. + ```mylam```: The mean of the marker scores of the diseased group after the Box-Cox transformation. + ```sylam```: The standard deviation of the marker scores of the diseased group after the Box-Cox transformation. + ```results```: A table that provides some indicative results: the AUC, the J (maximized Youden index), the estimated cutoff, the Sens, and the Spec along with their marginal CIs. + ```roc```: A function of the estimated Box-Cox ROC curve. You can use this to simply request TPR values for given FPR values. * **Example: ** ```{r, echo=TRUE, eval=TRUE} set.seed(123) x=rgamma(100, shape=2, rate = 8) y=rgamma(100, shape=2, rate = 4) scores=c(x,y) D=c(zeros(1,100), ones(1,100)) out=rocboxcox(marker=scores,D, 0.05, plots="on", printProgress=FALSE) summary(out) ``` --- **rocboxcoxCI** *** * **Description** + This function applies the Box-Cox transformation and carries out statistical inferences for the sensitivity at a given specificity (and vice versa). * **Usage** ```out=rocboxcoxCI(marker, D, givenSP=givenSP, givenSE=NA, alpha, plots) ``` ```out=rocboxcoxCI(marker, D, givenSP=NA, givenSE=givenSE, alpha, plots) ``` * **Arguments** + `marker`: A vector of length n that contains the biomarker scores of all individuals. + ```D```: A vector of length n that contains the true disease status of an individual, where 0 denotes a healthy/control individual, and 1 denotes a diseased individual. + ```givenSP```: A vector of specificity values that the user wants to fix/set, at which the sensitivity is to be estimated. In this case, the ‘givenSE’ argument needs to be set to NA. + ```givenSE```: A vector of sensitivity values that the user want to fix/set, at which the specificity is to be estimated. In this case, the ‘givenSP’ argument needs to be set to NA. + ```alpha```: Nominal level used to calculate the confidence intervals. A common choice is 0.05. + ```plots```: Valid inputs are "on" and "off". When set to "on", it returns both (1) the Box-Cox based ROC plot along with pointwise 95% confidence intervals for the full spectrum of FPRs and (2) a second plot that visualizes the confidence intervals at the given sensitivities or specificities. * **Value** + ```SPandCIs```: The specificity values and the CIs around them. + ```SEandCIs```: The specificity values and the CIs around them. + ```SEvalues```: The sensitivity values provided by the user at which the specificity was calculated. If the user did not provide any sensitivity values, this argument shoulb be set to NA. + ```SPvalues```: The specificity values provided by the user at which the sensitivity was calculated. If the user did not provide any specificity values, this argument should be set to NA. * **Example: ** ```{r, echo=TRUE, eval=TRUE} givenSP=c(0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9) givenSE=NA out=rocboxcoxCI(marker=scores ,D, givenSP=givenSP, givenSE=NA, alpha=0.05, plots="on") summary(out) ``` --- **checkboxcox2** *** * **Description** + This function tests whether the Box-Cox transformation is able to achieve approximate normality for your data. That is, it will allow the user to know whether it is appropriate to use all the methods discussed later on in this package. It is similar to the function checkboxcox but is designed for two markers instead of just one. * **Usage** ```checkboxcox2(marker1, marker2, D, plots, printShapiro = TRUE) ``` * **Arguments** + `marker1`: A vector of length n that contains the biomarker scores of all individuals for the first marker. + `marker2`: A vector of length n that contains the biomarker scores of all individuals for the second marker. + ```D```: A vector of length n that contains the true disease status of an individual. It is a binary vector containing 0 for the healthy/control individuals and 1 for the diseased individuals. + ```plots='on' or 'off'```: Valid inputs are "on" and "off". When set to "on", the user gets the histograms of the biomarker for both the healthy and the diseased group before and after the Box-Cox transformation for both marker1 and marker2. In addition, all eight corresponding qq-plots are provided. + ```printShapiro```: Boolean. When set to TRUE, the results of the Shapiro-Wilk test will be printed to the console. When set to FALSE, the results are suppressed. Default value is FALSE. * **Value** + ```res_shapiro```: A results table that contains the results of eight Shapiro-Wilk tests for normality testing. Four of these refer to normality testing of the healthy and the diseased groups before the Box-Cox transformation, and the remaining four refer to the Box-Cox transformed biomarkers scores for the healthy and the diseased groups. Thus, this testing process produces eight p-values. In addition, if the plots are set to 'on', then the output provides (1) the histograms of the biomarker for both the healthy and the diseased groups before and after the Box-Cox transformation for both marker1 and marker2, (2) all eight corresponding qq-plots, and (3) two plots (one for marker1, one for marker2) with the empirical ROC curve overplotted with the Box-Cox based ROC curve for visual comparison purposes. + ```transx1```: The Box-Cox transformed scores for the first marker and the healthy group. + ```transy1```: The Box-Cox transformed scores for the first marker and the diseased group. + ```transformation.parameter.1```: The estimated Box-Cox transformation parameter (lambda) for marker1. + ```transx2```: The Box-Cox transformed scores for the second marker and the healthy group. + ```transy2```: The Box-Cox transformed scores for the second marker and the diseased group. + ```transformation.parameter.2```: The estimated Box-Cox transformation parameter (lambda) for marker2. + ```pval_x1```: The p-value of the Shapiro Wilk test of normality for the marker1 healthy group (before the Box-Cox transformation). + ```pval_y1```: The p-value of the Shapiro Wilk test of normality for the marker1 diseased group (before the Box-Cox transformation). + ```pval_transx1```: The p-value of the Shapiro Wilk test of normality for the marker1 healthy group (after the Box-Cox transformation). + ```pval_transy1```: The p-value of the Shapiro Wilk test of normality for the marker1 diseased group (after the Box-Cox transformation). + ```pval_x2```: The p-value of the Shapiro Wilk test of normality for the marker2 healthy group (before the Box-Cox transformation). + ```pval_y2```: The p-value of the Shapiro Wilk test of normality for the marker2 diseased group (before the Box-Cox transformation). + ```pval_transx2```: The p-value of the Shapiro Wilk test of normality for the marker2 healthy group (after the Box-Cox transformation). + ```pval_transy2```: The p-value of the Shapiro Wilk test of normality for the marker2 diseased group (after the Box-Cox transformation). + ```roc1```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```roc2```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. * **Example: ** ```{r, echo=TRUE, eval=TRUE} #DATA GENERATION set.seed(123) nx <- 100 Sx <- matrix(c(1, 0.5, 0.5, 1), nrow = 2, ncol = 2) mux <- c(X = 10, Y = 12) X = mvtnorm::rmvnorm(nx, mean = mux, sigma = Sx) ny <- 100 Sy <- matrix(c(1.1, 0.6, 0.6, 1.1), nrow = 2, ncol = 2) muy <- c(X = 11, Y = 13.7) Y = mvtnorm::rmvnorm(ny, mean = muy, sigma = Sy) dx = pracma::zeros(nx,1) dy = pracma::ones(ny,1) markers = rbind(X,Y); marker1 = markers[,1] marker2 = markers[,2] D = c(rbind(dx,dy)) out=checkboxcox2(marker1, marker2, D, plots = "on") summary(out) ``` --- **comparebcAUC** *** * **Description** + This function provides a comparison of two correlated markers in terms of their AUCs (areas under the Box-Cox based ROC curves). Marker measurements are assumed to be taken on the same individuals for both markers. * **Usage** ```out=comparebcAUC(marker1, marker2, D, alpha, plots) ``` * **Arguments** + `marker1`: A vector of length n that contains the biomarker scores of all individuals for the first marker. + `marker2`: A vector of length n that contains the biomarker scores of all individuals for the second marker. + ```D```: A vector of length n that contains the true disease status of an individual. It is a binary vector containing 0 for the healthy/control individuals and 1 for the diseased individuals. + ```alpha```: Nominal level used to calculate the confidence intervals. A common choice is 0.05. + ```plots```: Valid inputs are "on" and "off". When set to "on" it returns the Box-Cox based ROC along with informative information about the two AUCs in the legend of the plot. * **Value** + ```resultstable```: A summary table of the comparison that contains the AUC of each marker, the p-value of the difference using the probit transformation, the p-value of the difference, and the confidence interval of the difference. + ```AUCmarker1```: The AUC of the first marker. + ```AUCmarker2```: The AUC of the second marker. + ```pvalue_probit_difference```: The p-value for the comparison of the AUCs. It employs the probit transformation that has greater power then when not using it (as opposed to the p-value provided in the 'pvalue difference' output argument provided later on). It is based on $Z^*=\frac{\Phi^{-1}\left(\hat{AUC}_{1}\right)-\Phi^{-1}\left(\hat{AUC}_{2}\right)}{\sqrt{Var\left(\Phi^{-1}\left(\hat{AUC}_{1}\right)\right)+Var\left(\Phi^{-1}\left(\hat{AUC}_{2}\right)\right)-2Cov\left(\Phi^{-1}\left(\hat{AUC}_{1}\right)),\Phi^{-1}\left(\hat{AUC}_{2}\right)\right)}}$. + ```pvalue_difference```: The p-value for the comparisons of the AUC (without the probit transformation). Simulations have shown that this is inferior to the 'pvalue probit difference'. This is based on $Z=\frac{\hat{AUC}_{1}-\hat{AUC}_{2}}{\sqrt{Var\left(\hat{AUC}_{1}\right)+Var\left(\hat{AUC}_{2}\right)-2Cov\left(\hat{AUC}_{1},\hat{AUC}_{2}\right)}}$. + ```CI_difference```: The confidence interval for the difference of the AUCs. It is based on $Z$ given above. + ```roc1```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```roc2```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```transx1```: The Box-Cox transformed scores for the first marker and the healthy group. + ```transy1```: The Box-Cox transformed scores for the first marker and the diseased group. + ```transformation.parameter.1```: The estimated Box-Cox transformation parameter (lambda) for marker 1. + ```transx2```: The Box-Cox transformed scores for the second marker and the healthy group. + ```transy2```: The Box-Cox transformed scores for the second marker and the diseased group. + ```transformation.parameter.2```: The estimated Box-Cox transformation parameter (lambda) for marker 2. * **Example: ** ```{r, echo=TRUE, eval=TRUE} #GENERATE SOME BIVARIATE DATA=== set.seed(123) nx <- 100 Sx <- matrix(c(1, 0.5, 0.5, 1), nrow = 2, ncol = 2) mux <- c(X = 10, Y = 12) X=rmvnorm(nx, mean = mux, sigma = Sx) ny <- 100 Sy <- matrix(c(1.1, 0.6, 0.6, 1.1), nrow = 2, ncol = 2) muy <- c(X = 11, Y = 13.7) Y=rmvnorm(ny, mean = muy, sigma = Sy) dx=zeros(nx,1) dy=ones(ny,1) markers=rbind(X,Y); marker1=markers[,1] marker2=markers[,2] D=c(rbind(dx,dy)) #==DATA HAVE BEEN GENERATED==== #===COMPARE THE AUCs of Marker 1 vs Maker 2 out=comparebcAUC(marker1, marker2, D, alpha=0.05, plots="on") summary(out) ``` --- **comparebcJ** *** * **Description** + This function provides a comparison of two correlated markers in terms of their J (Youden indices for Box-Cox based ROC curves). Markers measurements are assumed to be taken on the same individuals for both markers. * **Usage** ```out=comparebcJ(marker1, marker2, D, alpha, plots) ``` * **Arguments** + `marker1`: A vector of length n that contains the biomarker scores of all individuals for the first marker. + `marker2`: A vector of length n that contains the biomarker scores of all individuals for the second marker. + ```D```: A vector of length n that contains the true disease status of an individual, where 0 denotes a healthy/control individual, and 1 denotes a diseased individual. + ```alpha```: Nominal level used to calculate the confidence intervals. A common choice is 0.05. + ```plots```: Valid inputs are "on" and "off". When set to "on" it returns the Box-Cox based ROC along with informative information about the two AUCs in the legend of the plot. * **Value** + ```resultstable```: A summary table of the comparison that contains the maximized Youden Index of each marker, the p-value of the difference using the probit transformation, the p-value of the difference, and the confidence interval of the difference. + ```J1```: The maximized Youden Index (J) of the first marker. + ```J2```: The maximized Youden Index (J) of the second marker. + ```pvalue_probit_difference```: The p-value for the comparison of the Js. It employs the probit transformation that has greater power then when not using it (as opposed to the p-value provided in the 'pvalue difference' output argument provided later on). It is based on $Z^{*}=\frac{\hat{J}_{T2}-\hat{J}_{T1}}{\sqrt{Var(\hat{J}_{T1})+Var(\hat{J}_{T2})-2Cov(\hat{J}_{T1},\hat{J}_{T2})}}$ where $\hat{J}_{Ti}=\Phi^{-1}(\hat{J}_{i})$, $i=1,2$. + ```pvalue_difference```: The p-value for the comparisons of the J (without the probit transformation). Simulations have shown that this is inferior to the 'pvalue probit difference'. This is based on $Z=\frac{\hat{J}_{2}-\hat{J}_{1}}{\sqrt{Var(\hat{J}_{1})+Var(\hat{J}_{2})-2Cov(\hat{J}_{1},\hat{J}_{T})}}$. + ```CI_difference```: The confidence interval for the difference of the Js. This is based on $Z$ mentioned right above. + ```roc1```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```roc2```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```transx1```: The Box-Cox transformed scores for the first marker and the healthy group. + ```transy1```: The Box-Cox transformed scores for the first marker and the diseased group. + ```transformation.parameter.1```: The estimated Box-Cox transformation parameter (lambda) for marker 1. + ```transx2```: The Box-Cox transformed scores for the second marker and the healthy group. + ```transy2```: The Box-Cox transformed scores for the second marker and the diseased group. + ```transformation.parameter.2```: The estimated Box-Cox transformation parameter (lambda) for marker 2. * **Example: ** ```{r, echo=TRUE, eval=TRUE} #==DATA HAVE BEEN GENERATED==== #===COMPARE THE Js of Marker 1 vs Maker 2 out=comparebcJ(marker1, marker2, D, alpha=0.05, plots="on") summary(out) ``` --- **comparebcSens** *** * **Description** + This function provides a comparison of two correlated markers in terms of their sensitivities at a given specificity (for Box-Cox based ROC curves). Marker measurements are assumed to be taken on the same individuals for both markers. * **Usage** ```out=comparebcSens(marker1, marker2, D, alpha, atSpec, plots) ``` * **Arguments** + `marker1`: A vector of length n that contains the biomarker scores of all individuals for the first marker. + `marker2`: A vector of length n that contains the biomarker scores of all individuals for the second marker. + ```D```: A vector of length n that contains the true disease status of an individual, where 0 denotes a healthy/control individual, and 1 denotes a diseased individual. + ```alpha```: Nominal level used to calculate the confidence intervals. A common choice is 0.05. + ```atSpec```: The value of specificity at which the comparison of sensitivities will take place. + ```plots```: Valid inputs are "on" and "off". When set to "on" it returns the Box-Cox based ROC along with informative information about the two AUCs in the legend of the plot. * **Value** + ```resultstable```: A summary table of the comparison that contains the sensitivity of each marker at the given specificity, the p-value of the difference using the probit transformation, the p-value of the difference, and the confidence interval of the difference. + ```Sens1```: The sensitivity at the selected specificity for the first marker. + ```Sens2```: The sensitivity at the selected specificity for the second marker. + ```pvalue_probit_difference```: The p-value for the comparison of the sensitivities. It employs the probit transformation that has greater power then when not using it (as opposed to the p-value provided in the 'pvalue_difference' output argument provided later on). + ```pvalue_difference```: The p-value for the comparisons of the sensitivities (without the probit transformation). Simulations have shown that this is inferior to the 'pvalue_probit_difference'. + ```CI_difference```: The confidence interval for the difference of the sensitivities. + ```roc1```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```roc2```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```transx1```: The Box-Cox transformed scores for the first marker and the healthy group. + ```transy1```: The Box-Cox transformed scores for the first marker and the diseased group. + ```transformation.parameter.1```: The estimated Box-Cox transformation parameter (lambda) for marker 1. + ```transx2```: The Box-Cox transformed scores for the second marker and the healthy group. + ```transy2```: The Box-Cox transformed scores for the second marker and the diseased group. + ```transformation.parameter.2```: The estimated Box-Cox transformation parameter (lambda) for marker 2. * **Example: ** ```{r, echo=TRUE, eval=TRUE} out=comparebcSens(marker1=marker1, marker2=marker2, D=D, alpha =0.05, atSpec=0.8, plots="on") summary(out) ``` --- **comparebcSpec** *** * **Description** + This function provides a comparison of two correlated markers in terms of their specificities at a given sensitivity (for Box-Cox based ROC curves). Marker measurements are assumed to be taken on the same individuals for both markers. * **Usage** ```out=comparebcSpec(marker1, marker2, D, alpha, atSens, plots) ``` * **Arguments** + `marker1`: A vector of length n that contains the biomarker scores of all individuals for the first marker. + `marker2`: A vector of length n that contains the biomarker scores of all individuals for the second marker. + ```D```: A vector of length n that contains the true disease status of an individual, where 0 denotes a healthy/control individual, and 1 denotes a diseased individual. + ```alpha```: Nominal level used to calculate the confidence intervals. A common choice is 0.05. + ```atSens```: The value of sensitivity at which the comparison of specificities will take place. + ```plots```: Valid inputs are "on" and "off". When set to "on" it returns the Box-Cox based ROC along with informative information about the two AUCs in the legend of the plot. * **Value** + ```Spec1```: The specificity at the selected sensitivity for the first marker. + ```Spec2```: The specificity at the selected sensitivity for the second marker. + ```pvalue_probit_difference```: The p-value for the comparison of the specificities. It employs the probit transformation that has greater power then when not using it (as opposed to the p-value provided in the 'pvalue_difference' output argument provided later on). + ```pvalue_difference```: The p-value for the comparison of the specificities (without the probit transformation). Simulations have shown that this is inferior to the 'pvalue_probit_difference'. + ```CI_difference```: The confidence interval for the difference of the specificities. + ```roc1```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```roc2```: A function that refers to the ROC of the first marker. It allows the user to feed in FPR values and the corresponding TPR values. + ```transx1```: The Box-Cox transformed scores for the first marker and the healthy group. + ```transy1```: The Box-Cox transformed scores for the first marker and the diseased group. + ```transformation.parameter.1```: The estimated Box-Cox transformation parameter (lambda) for marker 1. + ```transx2```: The Box-Cox transformed scores for the second marker and the healthy group. + ```transy2```: The Box-Cox transformed scores for the second marker and the diseased group. + ```transformation.parameter.2```: The estimated Box-Cox transformation parameter (lambda) for marker 2. * **Example: ** ```{r, echo=TRUE, eval=TRUE} out=comparebcSpec(marker1=marker1, marker2=marker2, D=D, alpha =0.05, atSens=0.8, plots="on") summary(out) ``` --- **threerocs** *** * **Description** + This function provides a visual comparison of the Empirical ROC, the Box-Cox ROC, and the Metz binormal semi-parametric estimator of the ROC curve. It also computes the AUC for the curve corresponding to each method. * **Usage** ```threerocs2(marker1, marker2, D, plots) ``` * **Arguments** + `marker1`: A vector of length n that contains the biomarker scores of all individuals for the first marker. + `marker2`: A vector of length n that contains the biomarker scores of all individuals for the second marker. + ```D```: A vector of length n that contains the true disease status of an individual. It is a binary vector containing 0 for the healthy/control individuals and 1 for the diseased individuals. + ```plots='on' or 'off'```: Valid inputs are "on" and "off". When set to "on", the user gets a single plot containing the estimated ROC curves using the Empirical, Box-Cox, and Metz methods for each of the two provided markers. * **Value** + ```AUC_BoxCox1```: The AUC of the Box-Cox based ROC curve for the first marker. + ```AUC_BoxCox2```: The AUC of the Box-Cox based ROC curve for the second marker. + ```AUC_Metz1```: The AUC of the Metz binormal curve (as calculated by MRMCaov package using the "binormal" option) for the first marker. + ```AUC_Metz2```: The AUC of the Metz binormal curve (as calculated by MRMCaov package using the "binormal" option) for the second marker. + ```AUC_Empirical1```: The AUC of the empirical ROC curve for the first marker. + ```AUC_Empirical2```: The AUC of the empirical ROC curve for the second marker. * **Example: ** ```{r, echo=TRUE, eval=TRUE} out=threerocs2(marker1, marker2, D, plots = "on") summary(out) ```