function to train linear regression

LinearReg.R

## this function takes four inputs
## formula such as "mpg ~ wt + hp" (character)
## dataset
## train_size ratio between 0-1 (numeric)
## target such as "mpg" (character)

LinearRegression <- function(formula, data, train_size, target){
    ## split dataset into train and test  
    set.seed(99)
    n <- nrow(data)
    id <- sample(1:n, size = train_size*n, replace = FALSE)
    train_df <- data[id, ] 
    test_df <- data[-id, ]
    
    ## train model
    train_form <- as.formula(formula)
    lm_model <- lm(train_form, data = train_df)

    ## predictions
    train_p <- predict(lm_model)
    test_p <- predict(lm_model, newdata = test_df)

    ## evaluate and print result
    train_mae <- mean(abs(train_p - train_df[[target]]))
    test_mae <- mean(abs(test_p - test_df[[target]]))
    cat("Train MAE:", train_mae)
    cat("\nTest MAE:", test_mae)
}