[Limdep Nlogit List] Issues with LCM

[William Greene]

Dear Elizabeth.  It actually doesn't sound like the two class model worked
either.  There is nothing wrong with the program - other data sets and
models
work fine.  This does sound like the LCM may not be appropriate for your
data.
That can happen with any data set.  The failure to be able to estimate the
model
often suggests a mismatch between the model and the data.  You might try to
greatly simplify the model (temporarily) to see if it is possible to get
*any* form
of the model to converge properly.  If so, build outward from there.
Regards,
Bill Greene

On Tue, Nov 19, 2019 at 3:17 PM E Pienaar via Limdep <
limdep at mailman.sydney.edu.au> wrote:

> I am trying to run a latent class model with dummy coded data. There are 3
> programs per choice set.  Programs 2 and 3 have a cost associated with them
> ($10 to $100) and the other attributes are dummy coded.  Program 1 has no
> cost and the attributes are all zero.
>
> This is a large data set (22,370+ choice sets, i.e. over 61,000 lines of
> data).
>
> I am including an ASC in the model.
>
> I have written the code as follows:
> NLOGIT; Lhs = choice
> ; Choices = alta, altb, altc
> ; Rhs = ASC, ...., cost
> ; Alg = BFGS
> ; maxit = 200
> ; Tlg = 0.0001
> ; Tlb = 0.0001
> ; Tlf = 0.0001
> ; lcm
> ; pds = 3
> ; pts = 3
> ; parameters
> ; prb = class
> ; keep = pred $
>
> When I specify there should be 2 classes, the model converges but the p
> values for all variables in class 2 are 0.99 or above.
>
> If I run the model with 3 classes then I get the following error message:
> "Iterative procedure has converged
> Normal exit:  96 iterations. Status=0, F=    .1394539D+05
> Error   1027: Models - estimated variance matrix of estimates is singular"
>
> This suggests the model has converged but I also receive the error code
> claiming the variance matrix is singular.
>
> Thank you for any assistance in resolving this problem.
>
> Elizabeth
> _______________________________________________
> Limdep site list
> Limdep at mailman.sydney.edu.au
> http://limdep.itls.usyd.edu.au
>
>

-- 
William Greene
Department of Economics, emeritus
Stern School of Business, New York University
44 West 4 St.
New York, NY, 10012
URL: https://protect-au.mimecast.com/s/PWxCCP7yOZtYRMMWHzK1LJ?domain=people.stern.nyu.edu
Email: wgreene at stern.nyu.edu
Ph. +1.646.596.3296
Editor in Chief: Journal of Productivity Analysis
Editor in Chief: Foundations and Trends in Econometrics
Associate Editor: Economics Letters
Associate Editor: Journal of Business and Economic Statistics