Dear All,
is there anyone knowing solutions or references concerning the use of
non positive definite correlation matrices (e.g. coming from
tetrachoric correlation) in PCA and/or Factor analysis procedures ?
Thanks in advance.
*************************************
prof. Giuseppe BOARI
Istituto di Statistica
Universita' Cattolica del Sacro Cuore
via Necchi, 9
20123 MILANO
tel. +39-2-72342456
fax. +39-2-72342671
mailto:[log in to unmask]
*************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|