Original scientific paper

Determining the number of significant principal components on the basis of real variance of real and image variables intercorrelation matrix

Konstantin Momirović ; Inštitut za matematiko, fiziko in mehaniko Univerze v Ljubljani
Janez Štalec ; Inštitut za matematiko, fiziko in mehaniko Univerze v Ljubljani
Egon Zakrajšek ; Inštitut za matematiko, fiziko in mehaniko Univerze v Ljubljani

Abstract

It was suggested that a number of significant factors of complete intercorrelation matrixes could be determined to be just sufficient for explaining real variance of real and image variables intercorrelation matrix. The results of some empirical studies showed that such criterion was generally more superior than standard Guttman-Kaiser's criterion. In analysing matrixes of great hypothetical range the proposed criterion was also more superior that PB criterion by Momirović and Štalec.

Keywords

Hrčak ID:

240338

URI

https://hrcak.srce.hr/240338

Publication date:

1.1.1973.

Article data in other languages: croatian russian

Visits: 947 *