Building a stepwise regression model

In the absence of subject-matter expertise, stepwise regression can assist with the search for the most important predictors of the outcome of interest.

In this exercise, you will use a forward stepwise approach to add predictors to the model one-by-one until no additional benefit is seen. The donors dataset has been loaded for you.

This exercise is part of the course

Supervised Learning in R: Classification

View Course

Exercise instructions

Use the R formula interface with glm() to specify the base model with no predictors. Set the explanatory variable equal to 1.
Use the R formula interface again with glm() to specify the model with all predictors.
Apply step() to these models to perform forward stepwise regression. Set the first argument to null_model and set direction = "forward". This might take a while (up to 10 or 15 seconds) as your computer has to fit quite a few different models to perform stepwise selection.
Create a vector of predicted probabilities using the predict() function.
Plot the ROC curve with roc() and plot() and compute the AUC of the stepwise model with auc().

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

# Specify a null model with no predictors
null_model <- ___(___, data = ___, family = "___")

# Specify the full model using all of the potential predictors
full_model <- ___

# Use a forward stepwise algorithm to build a parsimonious model
step_model <- step(___, scope = list(lower = null_model, upper = full_model), direction = "___")

# Estimate the stepwise donation probability
step_prob <- ___

# Plot the ROC of the stepwise model
library(pROC)
ROC <- ___
plot(___, col = "red")
auc(___)

Edit and Run Code