[R] factanal AIC?

"Jens Oehlschlägel" oehl_list at gmx.de
Fri May 4 09:37:23 CEST 2007


Thanks  for answering. 

> your AIC is monotonic increasing

That was the reason I suspected something is wrong, but what?

Either the way I calculate the number of estimiated parameters?
Or the (scale of the) Likelihood itself?

> Have you tried a dimensional reduction technique or to investigate the wieghtings on your paramaters (eigenvalues in spectral decomp - principal 
components analysis).  Maybe some of those factors are not contributing?

The data was simulated, so we know what it is.

Best regards

Jens Oehlschlägel
"Feel free" - 10 GB Mailbox, 100 FreeSMS/Monat ...

More information about the R-help mailing list