Build a linear regression tree with multiple features
Predicting how effectively a drug will work for a given patient is one of the most practical and consequential challenges in modern healthcare. Clinicians often rely on broad population averages, yet the true response to a treatment varies widely based on factors such as age, sex, and dosage. As datasets grow richer and more personalised medicine becomes the norm, modelling these relationships with flexible, interpretable tools becomes essential. In this article, we’ll walk through how to build a linear regression tree—a model that combines the simplicity of linear regression with the segmentation power of decision trees—to better understand and predict drug effectiveness across diverse patient profiles.