Chap 10: Models that learn

Background

• In the previous chapter we introduced lm() and glm().
  • successfully handle the range of settings used in the course.
  • widely used in professional work

• We speak of training these models on the data.

• But there is also a notion of machine learning.

Some objectives for introducing tree models:

1. Provide insight into the model training process without requiring algebraic sophistication.
2. Suggest an answer to the question, “What’s the difference between training and learning?”
   • realizing when there is no difference between categories
   • picking out a small set of explanatory variables when there are a lot of possibilities.
3. Illustrate that there are more modeling techniques than linear and logistic models.
4. Show how models can be used for hypothesis generation.