Create a confusion matrix
As you saw in the video, a confusion matrix can be used to determine how well your model is performing. However, before creating the confusion matrix, you need to classify the predicted probabilities as 1 or 0, by using a cut-off.
Note: 1 means Inactive while 0 is Active.
prediction_test
, which contains the predicted probabilities of turnover for all cases in test_set
is available in your workspace.
This exercise is part of the course
HR Analytics: Predicting Employee Churn in R
Exercise instructions
- Turn the numeric predictions in
prediction_test
into a vector of categorical predictions using a cut-off of 0.5. - Create the confusion matrix using
prediction_categories
and actual values in the test set (test_set$turnover
).
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Classify predictions using a cut-off of 0.5
prediction_categories <- ___(prediction_test > 0.5, 1, 0)
# Construct a confusion matrix
conf_matrix <- ___(prediction_categories, test_set$turnover)
conf_matrix