PRINCIPAL COMPONENTS ANALYSIS
Daniel J. Denis · 2021
Principal components analysis attempts to reduce the dimensionality of a group of correlated variables into a set of mutually orthogonal linear combinations of the variables of lower dimension yet simultaneously attempting to explain most of the variance in the original variables. Harold Hotelling is historically recognized as an “inventor” of principal components analysis. In this chapter, the authors consider a summary of the formal development of principal components. They discuss the fact that each extracted eigenvector is associated with a respective eigenvalue. Each eigenvalue represents the variance for the given component. To demonstrate a very simple principal components analysis, the authors consider data featured in Pearson's 1901 paper. The scree plot is a graphical device used for helping to decide the number of worthwhile components to retain from a principal components or factor analysis. The authors compare the procedures of principal components analysis to that of least-squares regression.
Cited by 21 works. doi:10.1002/9781119583004.ch13