The structure of improper solutions in maximum likelihood factor analysis
The structure of maximum likelihood estimators in the Heywood case is investigated without any restrictions on parameters. It is shown that this structure is identical to that obtained by Jöreskog (1967) under the restriction that some uniquenesses are zero. As a result, the structure of generalized least-squares estimators is clarified.
| Year of publication: |
1987
|
|---|---|
| Authors: | Ihara, Masamori |
| Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 5.1987, 1, p. 35-37
|
| Publisher: |
Elsevier |
| Keywords: | AMS (1980) Subject Classification: 62H25 factor analysis improper solutions maximum likelihood estimators generalized east-squares estimators |
Saved in:
Saved in favorites
Similar items by person
-
Identification of inconsistent variates in factor analysis
Kano, Yutaka, (1994)
-
Ihara, Masamori, (1992)
-
A simple Marquardt algorithm for the nonlinear least-squares problem
Okamoto, Masashi, (1983)
- More ...