Train-Test Split

Learn how to split data into training and testing sets in R. This lecture covers essential techniques for preparing data for machine learning models.

Introduction

In this lecture, we will learn how to split data into training and testing sets in R. Splitting your data is a crucial step in machine learning, as it allows you to train your model on one subset of the data and evaluate its performance on another, ensuring that the model generalizes well to unseen data.

Key Concepts

1. Importance of Train-Test Split

The main reasons for splitting your data into training and testing sets are:

• Prevent Overfitting: By evaluating the model on a separate test set, you can check if the model overfits the training data.

• Model Validation: It helps in validating the model’s performance on unseen data, giving a better estimate of how the model will perform in real-world scenarios.

2. Proportion of Split

A common practice is to split the data into 70-80% for training and 20-30% for testing. The exact proportions can vary based on the size of your dataset and the specific use case.

Splitting Data in R

1. Using base R

You can use base R functions to split the data into training and testing sets.

# Sample data

set.seed(123)

data <- data.frame(

  x = rnorm(100),

  y = rnorm(100)

)

data$z <- 3 * data$x + 2 * data$y + rnorm(100)



# Splitting the data

train_indices <- sample(seq_len(nrow(data)), size = 0.7 * nrow(data))

train_data <- data[train_indices, ]

test_data <- data[-train_indices, ]



print(dim(train_data))  # Output: 70 3

print(dim(test_data))   # Output: 30 3

2. Using the caret Package

The caret package provides a convenient function createDataPartition() for splitting data.

# Loading the caret package

library(caret)



# Creating a sample dataset

set.seed(123)

data <- data.frame(

  x = rnorm(100),

  y = rnorm(100),

  z = 3 * rnorm(100) + 2 * rnorm(100) + rnorm(100)

)



# Splitting the data

trainIndex <- createDataPartition(data$z, p = 0.7, list = FALSE)

train_data <- data[trainIndex, ]

test_data <- data[-trainIndex, ]



print(dim(train_data))  # Output: 70 3

print(dim(test_data))   # Output: 30 3

3. Using the caTools Package

The caTools package provides the sample.split() function for splitting data.

# Installing and loading the caTools package

install.packages("caTools")

library(caTools)



# Splitting the data

set.seed(123)

split <- sample.split(data$z, SplitRatio = 0.7)

train_data <- subset(data, split == TRUE)

test_data <- subset(data, split == FALSE)



print(dim(train_data))  # Output: 70 3

print(dim(test_data))   # Output: 30 3

Example: Comprehensive Train-Test Split

Here’s an example of a comprehensive train-test split using the caret package.

# Loading the caret package

library(caret)



# Creating a sample dataset

set.seed(123)

data <- data.frame(

  x = rnorm(100),

  y = rnorm(100),

  z = 3 * rnorm(100) + 2 * rnorm(100) + rnorm(100)

)



# Splitting the data

trainIndex <- createDataPartition(data$z, p = 0.7, list = FALSE)

train_data <- data[trainIndex, ]

test_data <- data[-trainIndex, ]



# Training a simple linear model

model <- train(z ~ x + y, data = train_data, method = "lm")



# Making predictions on the test set

predictions <- predict(model, newdata = test_data)



# Evaluating the model's performance

mse <- mean((test_data$z - predictions)^2)

print(paste("Mean Squared Error:", mse))

Summary

In this lecture, we covered how to split data into training and testing sets in R using base R functions, the caret package, and the caTools package. Splitting your data is a crucial step in building machine learning models, ensuring that they generalize well to unseen data.

Further Reading

For more detailed information, consider exploring the following resources:

• Train-Test Split in R

• R Documentation on Data Splitting

• caret Package Documentation

Call to Action

If you found this lecture helpful, make sure to check out the other lectures in the ML R series. Happy coding!