Some applications of principal component analysis to well-log data

Abstract

Principal component analysis is a multivariate statistical technique used to ascertain the effective quantity of information existing in a data set. The procedure for converting original data into principal components entails the standardization of these data and subsequent calculation of the correlation matrix, its eigenvectors and eigenvalues (used to obtain the coefficients of transformation), and the proportion of total variance associated with each principal component. Some applications of this methodology include facies determination (in conjunction with other methods, such as discriminant analysis); preliminary porosity estimation; and well-to-well correlation.