Train a model and generate class predictions

This function takes a model specification and a recipe, fits the model on the training data, and returns predictions on the testing data.

Usage

train_and_predict(model_spec, data_train, data_test, recipe)

Arguments

model_spec

A tidymodels model specification.

data_train

A data frame containing the training data.

data_test

A data frame containing the testing data.

recipe

A tidymodels recipe for preprocessing the data.

Value

A data frame containing the predicted classes along with the original data_test columns.

Examples

# Example usage of the `train_and_predict` function
library(testthat)
library(parsnip)
library(recipes)
library(workflows)
library(dplyr)
library(rsample)
set.seed(123)
data <- tibble(
  class = sample(c("A", "B"), 150, replace = TRUE),
  feature1 = rnorm(150),
  feature2 = rnorm(150)
)

# Split the data into training and testing sets
data_split <- initial_split(data, prop = 0.8)
data_train <- training(data_split)
data_test <- testing(data_split)

# Define the model specification
model_spec <- decision_tree() %>%
  set_engine("rpart") %>%
  set_mode("classification")

# Define the recipe (preprocessing)
recipe <- recipe(class ~ feature1 + feature2, data = data_train) %>%
  step_normalize(all_numeric())

# Run the function
predictions <- train_and_predict(model_spec, data_train, data_test, recipe)
head(predictions)
#> # A tibble: 6 × 4
#>   .pred_class class feature1 feature2
#>   <fct>       <chr>    <dbl>    <dbl>
#> 1 B           A        0.994   -0.789
#> 2 A           B        0.548   -0.502
#> 3 B           A       -0.600    1.90 
#> 4 B           B        2.19    -0.101
#> 5 B           A       -0.710   -0.376
#> 6 A           B        0.257   -0.562