Manual regularization with Lasso
The attrition
dataset has 30 variables. Your Human Resources department asks you to build a model that is easy to interpret and maintain. They specifically want to reduce the number of features so that your model is as interpretable as possible.
In this exercise, you'll use Lasso to reduce the number of variables in your model automatically. In this first attempt, you will manually input a penalty and observe the model's behavior.
train
and test
data, and a basic recipe
are already loaded for you.
This exercise is part of the course
Feature Engineering in R
Exercise instructions
- Set your logistic regression model to use the
glmnet
engine. - Set arguments to run Lasso with a penalty of 0.06.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
model_lasso_manual <- logistic_reg() %>%
# Set the glmnet engine for your logistic regression model
___(___) %>%
# Set arguments to run Lasso with a penalty of 0.06
set_args(mixture = ___, ___ = ___)
workflow_lasso_manual <- workflow() %>%
add_model(model_lasso_manual) %>%
add_recipe(recipe)
fit_lasso_manual <- workflow_lasso_manual %>%
fit(train)
tidy(fit_lasso_manual)