A User's Guide to Principal Components

TLDR

Principal component analysis, developed by Pearson and Hotelling in the early 20th century, remains a widely studied multivariate technique with ongoing interest reflected in recent literature. The book serves as a practical guide for data analysts and a comprehensive research resource on PCA. The book focuses on applied examples of PCA, providing extensive applications while omitting detailed proofs. Jackson’s book significantly expands the PCA literature and is likely the most comprehensive text currently available. The abstract includes an incomplete other section.

Abstract

Although confirmatory is now heavily used among educational and psychological researchers, interest in principal component (PCA) as a distinct method has continued. PCA is one of the oldest multivariate statistical techniques. The basic ideas were set forth by Karl Pearson around the turn of the century. Harold Hotelling developed PCA more fully in a 1933 paper in the Journal of Educational Psychology. Three recent books (Dunteman, 1989; Flury, 1988; Jolliffe, 1986), and a recent issue of Multivariate Behavioral Research devoted to the factor vs. component analysis debate (Velicer & Jackson, 1990), all attest to the continuing interest in PCA. The new book by J. Edward Jackson adds substantially to the literature on the topic, and may be the most comprehensive text available at present. Intended as a guide for the data analyst, it is also very useful as a source for research and literature on PCA; the reference list contains over 1,000 citations. Applications are emphasized throughout the text, with rigorous proofs generally absent. Readers who have had a graduate level course in multivariate statistics will be able to absorb most of the text. Two brief