Train a regression tree
As you know already, decision trees are a useful tool for classification problems. Moreover, you can also use them to model regression problems. The structural difference is that there will be numeric values (instead of classes) on the leaf nodes.
In this exercise, you will use the chocolate dataset to fit a regression tree. This is very similar to what you already did in Chapter 1 with the diabetes
dataset.
Available in your workspace is the training data chocolate_train
.
This exercise is part of the course
Machine Learning with Tree-Based Models in R
Exercise instructions
- Build
model_spec
, a regression tree specification. - Using the
chocolate_train
data frame, usemodel_spec
to train a regression tree that predictsfinal_grade
using only the numerical predictors in the data.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
library(tidymodels)
# Build the specification
model_spec <- decision_tree() %>%
set_mode(___) %>%
set_engine(___)
# Fit to the data
model_fit <- model_spec %>%
___(formula = ___,
data = ___)
model_fit