Predicting adult income with decision-tree-based models

ML Studio comes with three decision-tree-based algorithms for two-class classification: the Two-Class Decision Forest, Two-Class Boosted Decision Tree, and Two-Class Decision Jungle modules. These are known as ensemble models where more than one decision trees are assembled to obtain better predictive performance. Though all the three are based on decision trees, their underlying algorithms differ.

We will first build a model with the Two-Class Decision Forest module and then compare it with the Two-Class Boosted Decision Tree module for the Adult Census Income Binary Classification dataset module, which is one of the sample datasets available in ML Studio. The dataset is a subset of the 1994 US census database and contains the demographic information of working adults over the 16 years age limit. Each instance or example in the dataset has a label or class variable that states whether a person earns 50K a year or not.

Create an experiment...