Topic: Rethinking Principal Component Analysis
Abstract: Principal Component Analysis (PCA) is an essential tool for data scientists used both as an analysis tool in its own right and as a component of more complex analysis pipelines. PCA is often introduced as a decomposition of the variance matrix. In this talk, I argue that a regression approach is a more informative formulation that leads to deeper understanding of both the good properties of the method for data summary and analysis and of appropriate generalizations.