Correlation and compatibility in multivariate data analysis

Eigenvectors reflect active processes (real principal components or factors). The concepts of inherent compatibility and global compatibility are introduced. The inherent compatibility coefficient is used to study individual processes, whereas the global compatibility coefficient is used to study the combined nature and effect of each of the processes. The inherent and global variable compatibility of triterpanes is derived from a data base of 216 oil samples, and important pathways of the formation of oil from sedimentary kerogen are elucidated. The inefficiency of conventional data analysis is demonstrated.