Data Science Essentials in Python

Unit 51

Surviving in Random Decision Forests

A decision tree is a supervised machine learning tool. It uses a tree-style graph where each node contains a test on a certain data set attribute, and the branches incident to the node correspond to the outcomes of the test. If you use trees, you have to train them before use. Training consists of presenting various predictors and corresponding labels (features) to the tree and adjusting the node test conditions accordingly. (Surely, you don’t do training by hand!)

A random decision forest regressor uses a number (an ensemble) of classifying decision trees on various sub-samples of the data set and averages the predictions to improve accuracy. The module sklearn.ensemble provides the constructor ...