[R] factanal doesn't work with singular covariance matrix?

[Agus Salim]

Dear R-users,

I recently used factanal function in mva library to perform factor analysis.
I notice that the function will give an error message that the log likelihood
function is infinite whenever the observed covariance matrix is singular.
This is because the function uses multivariate normal log-likelihood which
involve log determinant of the observed covariance matrix.

But I wonder if the inclusion of log determinant of the observed covariance
matrix is necessary here because it doesn't involve any parameters we
want to estimate, so we can consider it as constant term.

I have tried two objective functions with and without the log determinant term
and they seem to give the same result.

What do you think?

Thanks,
Agus

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._