---
title: "Model Configurations"
author: 
  - name: Shixiang Wang
    affiliation: Central South University
    email: wangshx@csu.edu.cn
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Model Configurations}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  out.width = "100%"
)
```

```{r setup}
library(bregr)
```

## Interaction Terms

The **bregr** package provides comprehensive support for the processing and visualization of interaction terms in regression models, provided that the interaction term is inherently supported by the model.

Below is an illustrative example using a linear model:

```{r}
rv <- br_pipeline(
  data = mtcars,
  y = "mpg", x = c("cyl", "disp*hp"), x2 = "am",
  method = "lm"
)
```

Examine the results:

```{r}
br_get_results(rv, tidy = TRUE)
```

Visualize the results:

```{r fig.width=8, fig.height=4, dpi=150}
br_show_forest(rv)
```

Additionally, specified models can be visualized using the functions provided by **bregr**:

```{r fig.width=7, fig.height=3}
br_show_forest_ggstats(rv, idx = 2)
```

Or compare them:

```{r fig.width=7, fig.height=4}
br_show_forest_ggstats(rv)
```

## Customized GLM Family

For GLM models, the family object can be further customized with various arguments. By default, **bregr** directly supports GLM family functions with their default arguments, which can be viewed using:\

```{r}
br_avail_methods()
```

However, for alternative GLM configurations, modifications can be made directly. For instance, consider the following example using a quasi-family with specified variance and link functions:

```{r}
data <- data.frame(
  x = rnorm(100)
)
data$y <- rpois(100, exp(1 + data$x))
head(data)
```

```{r}
rv <- br_pipeline(
  data = data,
  y = "y", x = "x",
  method = 'quasi(variance = "mu", link = "log")'
)
br_get_results(rv, tidy = TRUE)
```

This is equivalent to:

```{r}
glm(y ~ x, data = data, family = quasi(variance = "mu", link = "log")) |>
  summary()
```