---
title: "multinom models"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{multinom models}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include = FALSE}
if (requireNamespace("nnet", quietly = TRUE)) {
  library(tidypredict)
  library(nnet)
  library(dplyr)
  eval_code <- TRUE
} else {
  eval_code <- FALSE
}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = eval_code
)
```

| Function                                                      |Works|
|---------------------------------------------------------------|-----|
|`tidypredict_fit()`, `tidypredict_sql()`, `parse_model()`      |  ✔  |
|`tidypredict_to_column()`                                      |  ✗  |
|`tidypredict_test()`                                           |  ✗  |
|`tidypredict_interval()`, `tidypredict_sql_interval()`         |  ✗  |
|`parsnip`                                                      |  ✔  |

`nnet::multinom()` fits multinomial log-linear models. Because these models
predict one probability per outcome class, `tidypredict_fit()` returns a *named
list* of expressions, one for each class, rather than a single expression. The
expressions implement the softmax over the per-class linear predictors, with the
first level of the outcome acting as the reference class.

Since the output is a list, `tidypredict_to_column()` and `tidypredict_test()`
are not supported.

## `tidypredict_` functions

```{r}
library(nnet)

model <- multinom(Species ~ ., data = iris, trace = FALSE)
```

- Create the R formulas, one per class
    ```{r}
fit <- tidypredict_fit(model)
names(fit)
fit[["setosa"]]
    ```

- Add the predictions to the original table
    ```{r}
library(dplyr)

iris %>%
  mutate(!!!tidypredict_fit(model)) %>%
  glimpse()
    ```

- Confirm that the results match the model's `predict()` results
    ```{r}
probs <- sapply(fit, \(f) rlang::eval_tidy(f, iris))
all.equal(unname(probs), unname(predict(model, iris, type = "probs")))
    ```

## parsnip

`parsnip` fitted models are also supported by `tidypredict`:
```{r}
library(parsnip)

p_model <- multinom_reg() %>%
  set_engine("nnet") %>%
  fit(Species ~ ., data = iris)
```

```{r}
tidypredict_fit(p_model)[["virginica"]]
```

## Parse model spec

Here is an example of the model spec:
```{r}
pm <- parse_model(model)
str(pm, 2)
```