Principal Components Analysis

Daniel J. Denis · 2018

Principal components analysis (PCA) is a data reduction technique useful for summarizing or describing the variance in a set of variables into fewer dimensions than there are variables in that data set. In SPSS, PCA is given as an “option” under the general name of factor analysis, even though the two procedures are distinct. To demonstrate the technique of principal components using SPSS, this chapter provides a very easy example from Karl Pearson's original 1901 data on a covariance matrix of only two variables, and then demonstrates a more realistic PCA on a correlation matrix of many more variables. To get component scores on each principal component, the researchers can use SPSS's automated feature to compute factor scores for them. SPSS further allows them to produce what are called loading plots of the two components plotted against each other.

Cited by 6 works. doi:10.1002/9781119465775.ch12

Authors