Q-mode factor analysis of compositional data

Computers & Geosciences



An extended form of the method of Q-mode factor analysis may be used if the rows of the data matrix sum to a constant. The constant row-sum provides a means for scaling the factor model to conform to the units of the original data in such forms as proportions, percentages, or parts per million. The scale factors for an m-factor model are used to derive composition scores and loadings, which lead to the same reproduced data regardless of the choice of reference axes (end-members). Factor-variance diagrams may be constructed to show the proportion of the variance that will be accounted for in each compositional variable by models containing various numbers of end-members. The method may be used to develop a petrologic mixing model by searching for end-members that are geologically plausible. ?? 1976.

Additional publication details

Publication type Article
Publication Subtype Journal Article
Title Q-mode factor analysis of compositional data
Series title Computers & Geosciences
DOI 10.1016/0098-3004(76)90003-0
Volume 1
Issue 3
Year Published 1976
Language English
Publisher Elsevier
Publisher location Amsterdam, Netherlands
Larger Work Type Article
Larger Work Subtype Journal Article
Larger Work Title Computers and Geosciences
First page 147
Last page 159