Exercise

Generate a tuning grid

The standard hyperparameters of most models provide a good fit for most datasets. Yet, they need optimization for the best performance. Otherwise, it's like driving a car with an activated hand brake. Release the brake and tune your models!

In this exercise, you'll create two objects that serve as a starting point: a tuning grid (a set of hyperparameter combinations) and a model specification that you later train with every value of the grid.

Instructions 1/2

undefined XP

1

2

Create an rpart powered classification tree specification, flagging the tree_depth and cost_complexity parameters for tuning.

.css-6su6fj{-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;}Exercise

Instructions 1/2

Exercise