Introduction

Multivariate data matrices analyzed by principal component analysis (PCA) are often accompanied by auxiliary information about the rows and columns of the data matrices. For example, the rows of a data matrix may represent subjects for whom some demographic information (e.g., gender, age, level of education, etc.) is available. The columns may represent stimuli constructed by manipulating several attributes with known values. Constrained principal component analysis (CPCA) incorporates such information in PCA of the main data matrix.