Regression Models

This part explains how to model relationships between variables for explanation and prediction. It moves from linear models to generalized models and tree-based methods, with emphasis on interpretation, diagnostics, and practical model comparison.

The workflow is model-first and diagnostics-driven: fit a candidate model, inspect residual behavior and assumptions, compare alternatives, and keep only models that are both interpretable and empirically defensible.

How This Part Is Structured

Block	Purpose	Main chapters
Linear models	Explain continuous outcomes using one or more predictors	Simple Linear Regression Model, Multiple Linear Regression Model
Generalized and categorical outcomes	Model non-Gaussian outcomes and categorical classes	Logistic Regression, Generalized Linear Model, Multinomial/Ordinal Models
Time-to-event and tree-based modeling	Extend to survival data and nonparametric partitioning	Cox Proportional Hazards, Conditional Inference Trees
Practical diagnostics	Connect estimation to assumption checks and test logic	Hypothesis Testing with Linear Regression Models from a Practical Point of View
Integrated practice	Consolidate regression concepts	Problems

How To Use It

If your target variable is continuous, start with SLR/MLR.
If your target is binary or categorical, start with Logistic/GLM or Multinomial/Ordinal models.
If interpretability with simple decision rules is key, evaluate trees early.

Use Appendix A — Method Selection Guide (Appendix A) for quick routing and revisit descriptive diagnostics whenever model assumptions are unclear.