Predictions of a single model
To calculate the performance of a classification model you need to compare the actual values of Attrition
to those predicted by the model.
When calculating metrics for binary classification tasks (such as precision and recall), the actual and predicted vectors must be converted to binary values.
In this exercise, you will learn how to prepare these vectors using the model and validate data frames from the first cross-validation fold as an example.
This exercise is part of the course
Machine Learning in the Tidyverse
Exercise instructions
- Extract the
model
and thevalidate
data frame from the first fold of the cross-validation. - Extract the
Attrition
column from thevalidate
data frame and convert the values to binary (TRUE/FALSE). - Use
model
to predict the probabilities of attrition for thevalidate
data frame. - Convert the predicted probabilities to a binary vector, assume all probabilities greater than
0.5
are TRUE.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Extract the first model and validate
model <- cv_models_lr$___[[___]]
validate <- cv_models_lr$___[[___]]
# Prepare binary vector of actual Attrition values in validate
validate_actual <- ___ == "Yes"
# Predict the probabilities for the observations in validate
validate_prob <- predict(___, ___, type = "response")
# Prepare binary vector of predicted Attrition values for validate
validate_predicted <- ___