Identifying models of factor analysis and using of generalized inverses

ABSTRACT PhD. THESIS IDENTIFYING MODELS OF FACTOR ANALYSIS AND USING OF GENERALIZED INVERSES Seval SÜZÜLMÜŞ DEPARTMENT OF STATISTICS INSTITUTE OF NATURAL AND APPLD2D SCIENCES UNIVERSITY OF ÇUKUROVA Supervisor: ProfDr. Sadullah SAKALLIO?LU Year: 2005, Pages: 145 Jury: ProfDr. Sadullah SAKALLIO?LU Prof.Dr. Fikri AKDENİZ ProfDr. H. Altan ÇABUK Prof.Dr. İhsan ÜNVER Doç.Dr. Selahattin KAÇIRANLAR In this thesis examined identifying of factor analysis model and the similarities and differences between factor analysis and principal component analysis. In factor analysis model for estimation parameters Principal Component Analysis, Maxsimum Likelihood Analysis, Unweighted Least Squares and Generalized Least Squares methods are used. In these methods the variance covariance matrice £ is positive definite. In a few methods, when S is not positive definite generalized inverse is used. So we investigate the case of the matrix £ is nonnegative-definite matrix. Key Words: : Factor Analysis, Principal Component Analysis, Rotation of Factors, Generalized Inverse. n