Application of a coherent multi-dimensional technique to the central part of the Mexican Volcanic Belt

Abstract

A combination of logratio transformation (additive logratio) and linear discriminant analysis (LDA) is a powerful experimental data-handling coherent multi-dimensional probabilitybased statistical technique, which is superior to the use of conventional bi-variate or ternary diagrams. Application of this procedure is illustrated with three subsets of compositional data for 2270 igneous rock samples from the central part of the Mexican Volcanic Belt (MVB). A combination of isometric logratio transformation and LDA will be used in future to test these and other sub-compositions on an updated worldwide compilation of geochemical data, which will enable us to achieve newer sets of more versatile, efficient, and coherent diagrams and related probability estimates.