added 'dimnames' argument to ML and FIML objective functions

5 replies [Last post]
mspiegel's picture
Offline
Joined: 07/31/2009

We added an optional 'dimnames' argument to the functions mxFIMLObjective() and mxMLObjective(). The argument expects a vector of character, such as c('foo', 'bar', 'baz'). The objective function uses this vector to populate the column names of the means vector, and the row and column names of the covariance matrix. With the 'dimnames' argument it is no longer necessary to explicitly specify the dimnames of the means and covariances, although it is still possible. If dimnames are specified for the means or covariances matrices, and the 'dimnames' argument is used by the objective function, and the dimnames values are not identical, then an error is thrown.

The demos and Sphinx documentation has been updated as well.

mspiegel's picture
Offline
Joined: 07/31/2009
PS: Those dimnames are not

PS: Those dimnames are not applied in the summary.

Fixed in r911.

tbates's picture
Offline
Joined: 07/31/2009
mxMLObjective ? code doesn't

mxMLObjective ? code doesn't run: complains about no dimnames.

Is this the fix?

A <- mxMatrix(values = 0.5, nrow = 2, ncol = 1,
free = TRUE, name = "A")

D <- mxMatrix(type = "Diag", values = c(0, 0.5),
free = c(FALSE, TRUE), nrow = 2, name = "D")

expectedCov <- mxAlgebra(A %*% t(A) + D, "expectedCov")
observedCov <- mxData(matrix(c(1.2, 0.8, 0.8, 1.3), nrow = 2, ncol = 2), 'cov', numObs = 150)
objective <- mxMLObjective(covariance = "expectedCov", dimnames=c('x','y'))
model <- mxModel("mxMLObjective example", A, D, expectedCov, objective, observedCov)

summary(mxRun(model))

PS: Those dimnames are not applied in the summary.

mspiegel's picture
Offline
Joined: 07/31/2009
Fixed in r910

Fixed in r910

tbates's picture
Offline
Joined: 07/31/2009
that's a nice feature

that's a nice feature michael.

I wonder if mxRun could be more specific when if finds an un-dimensioned mxMLObjective?

> fit <- mxRun(model)
Running ACE
Error: The expected covariance matrix associated with the ML objective in model 'ACE' does not contain dimnames.

nb: The expected covariance matrix which lacks dimnames is in a submodel of "ACE" called "modelMZ", and "ACE" uses an mxAlgebraObjective.... so a more helpful error would say "Error: The expected covariance matrix associated with the ML objective in 'ACE' sub-model "modelMZ" does not contain dimnames."

mspiegel's picture
Offline
Joined: 07/31/2009
I wonder if mxRun could be

I wonder if mxRun could be more specific when if finds an un-dimensioned mxMLObjective?

Fixed in revision 912.