Re: Getting parameter estimates for nominal logistic regression

[Steve Hodge <steve@nmr.mgh.harvard.edu>]

I'm sure someone has an answer to your question (why doesn't "Get 
Estimates" work in a nominal logistic model?), but in the meantime here's 
a work-around:

 FM=Fit Model(
        Y( :Y),
        Effects( :X1,  :X2,  :X3,  :X4),
        Personality(Nominal Logistic)
 );

 Fit = FM << Run Model(Save Probability Formula);
 rFit = Fit << get report;
 B = rFit["Parameter Estimates"][columnbox("Estimate")] << get as 
matrix;
(the estimates are now in "B")


Steve





On Sat, 12 Mar 2005, James Kern wrote:

> I am scripting the Fit Model (Nominal Logistic) platform and would like to 
> get the parameter estimates into a table.  Here's what I've tried:
> FM=Fit Model(Y( :Y), Effects( :X1,  :X2,  :X3,  :X4), Personality(Nominal 
> Logistic));
> Fit = FM<<Run Model(Save Probability Formula);
> B = Fit<<Get Estimates;
>
> This approach works for the Standard Least Squares platform.  The parameters 
> end up in a matrix called "B" that I can then send to a table.  But, when I 
> use the Nominal Logistic platform the matrix is empty.
> Any help is greatly appreciated. - Jim