Projection Pursuit Regression: Nonlinear Modeling and Interpretability

This lecture introduces Projection Pursuit Regression, a method inspired by tomography, for modeling nonlinear relationships in data. It discusses the decomposition of functions into smooth components and the challenges of interpretability when embracing nonlinearity. The instructor explains the fitting process, the trade-offs between interpretability and complexity, and the use of neural networks for regression. The lecture concludes with a discussion on the computational aspects of fitting complex models and the implications for data interpretation.