15.6 Feature Interpretation

• We can use the vip package to create a graph of feature importance

library(vip)
vip(cv_tree1, num_features = 20, geom = "point")

Figure 15.3: Variable importance based on the total reduction in MSE