Calculate cross-validated performance

It is crucial to optimize models using a carefully selected metric aimed at achieving the goal of the model.

Imagine that in this case you want to use this model to identify employees that are predicted to leave the company. Ideally, you want a model that can capture as many of the ready-to-leave employees as possible so that you can intervene. The corresponding metric that captures this is the recall metric. As such, you will exclusively use recall to optimize and select your models.

This exercise is part of the course

Machine Learning in the Tidyverse

View Course

Exercise instructions

Calculate the recall by comparing the actual with the predicted responses for each fold and assign it to the validate_recall column.
Print the validate_recall column.
Print the mean of this column.

Hands-on interactive exercise

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

# Calculate the validate recall for each cross validation fold
cv_perf_recall <- cv_prep_lr %>% 
  mutate(validate_recall = map2_dbl(___, ___, 
                                    ~recall(actual = .x, predicted = .y)))

# Print the validate_recall column
cv_perf_recall$___

# Calculate the average of the validate_recall column
___

Edit and Run Code