We introduce two mathematical propositions which will be developed for constructing a method for qualitative variables selection based on RV coefficient in the frame of Multiple Correspondence Analysis. Based on those propositions, variables selection will be transformed to the selection of set of Escoufierâs operators where its average maximizes the RV coefficient. The result is very promising. We will compare it to Escoufierâs method1).
Copyrights © 1998