The Wonderful World of ML #5, Trees - There's Power in Numbers

Trees are conceptually simple models, but they really shine for their wonderful interpretability. In other words, they are easy to explain and you don't need to be math wizard to understand them.

By themselves however, trees aren't as powerful as most other machine learning algorithms. However, if they are ensembled together properly, their collective performance can be quite impressive.

Though most typically used for classification, trees can also be used for regression. In this session, we'll touch on regression trees briefly, but focus primarily on classification. We'll build some trees from scratch and then spend some time discussing the 3 common methods for ensembling them: bagging, random forest, and boosting.