Supervised Learning in R: Regression
Nina Zumel and John Mount
Win-Vector, LLC
Rules of the form:
Non-linear concepts
non-additive interactions
Pro: Trees Have an Expressive Concept Space
Model | RMSE |
---|---|
linear | 0.1200419 |
tree | 0.1072732 |
Con: Coarse-Grained Predictions
Trees Predict Axis-Aligned Regions
Each color is a different predicted value
It's Hard to Express Lines with Steps
Ensembles Give Finer-grained Predictions than Single Trees
Ensemble Model Fits Animal Intelligence Data Better than Single Tree
Model | RMSE |
---|---|
linear | 0.1200419 |
tree | 0.1072732 |
random forest | 0.0901681 |
Supervised Learning in R: Regression