Re: st: Finite mixture model (FMM) estimation of probability determinants

[Federico Belotti <f.belotti@gmail.com>]

You can estimate models sequentially (as the help file reports, this is the preferred estimation approach)

fmm y x1 x2 x3 x4, comp(2) mix(normal)
fmm y x1 x2 x3 x4, comp(2) mix(normal) prob(x1 x2 x3 x4)

But in this way  "-fmm- will not check for reasonableness of the prior -fmm- estimates, so proceed carefully.  Otherwise, -fmm- expects the user to specify starting values using the -from(init_specs)- option", that is

fmm y x1 x2 x3 x4, comp(2) mix(normal) prob(x1 x2 x3 x4) from(b0)

where b0 is an appropriate vector of starting values of the same dimension of the unknown parameter vector to be estimated.

There are no technical restrictions on the -prob()- variable list.

Hope this helps,
Federico

Inviato da iPad

Il giorno 05/ago/2013, alle ore 21:39, "Julia Frances Slejko" <slejko@uw.edu> ha scritto:

> Hi,
> I am using FMM to estimate a mixture of two normal distributions.  I have
> been able to get the parameter estimates and the posterior probabilities
> themselves.  Now, I would like to use the Model 2 statement cited in the FMM
> help doc to estimate the determinants of the posterior probabilities.
> I have used the syntax as described in FMM help:
> 
> fmm y x1 x2 x3 x4, comp(2) mix(normal) prob(x1 x2 x3 x4)
> 
> But I get the following error: 'provide starting values or estimate 2
> component Normal model with constant component probabilities'.
> 
> Must I do this in two steps?  Is it OK for both var lists to be identical?
> The FMM help does not address starting values so I am not sure how to do
> this.
> 
> Any suggestions?
> 
> Thank you,
> Julia
> --
> Julia F. Slejko, PhD
> Postdoctoral Fellow
> 
> Pharmaceutical Outcomes Research and Policy Program
> School of Pharmacy | University of Washington
> Box 357630 | Seattle, WA 98195-7630
> mobile 720.937.9605
> fax 206.543.3835
> slejko@uw.edu
> 
> 
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/