Thanks a lot! - it did work almost exactly as you suggest, except that, at least in interactive mode, I had to specify the init(b0) option after
-ml model ..., init(b0)-
rather than
-ml maximize, init(b0)-
Anyway, now I know how to proceed - thanks again for all suggestions.
>-----Original Message-----
>From: [email protected]
>[mailto:[email protected]]On Behalf Of
>Austin Nichols
>Sent: Wednesday, August 29, 2007 6:04 PM
>To: [email protected]
>Subject: [SPAM] - Re: [SPAM] - Re: st: Nested maximum likelihood
>algorithms - Email has different SMTP TO: and MIME TO: fields in the
>email addresses - Email has different SMTP TO: and MIME TO: fields in
>the email addresses
>
>
>��� ���� <[email protected]>:
>A call to -ml maximize- can include a init(b0) option to specify a
>starting point and an iterate(1) option to specify iterating only
>once. The resulting b vector could be passed to another program,
>transformed to a new b0 and then passed to another call of -ml
>maximize, init(b0) iterate(1)- ad infinitum (or something short of
>infinitum, perhaps).
>
>On 8/29/07, ��� ���� <[email protected]> wrote:
>> At first attempt that does not seem to work, but thanks a
>lot for a useful hint!
>>
>> Alexis
>>
>> >-----Original Message-----
>> >From: [email protected]
>> >[mailto:[email protected]]On Behalf Of
>> >Scott Merryman
>> >Sent: Wednesday, August 29, 2007 3:09 PM
>> >To: [email protected]
>> >Subject: [SPAM] - Re: st: Nested maximum likelihood
>algorithms - Email
>> >has different SMTP TO: and MIME TO: fields in the email addresses
>> >
>> >
>> >Could you set the maximum number of iterations for the ML code to 1.
>> >
>> >Scott
>> >
>> >On 8/29/07, ��� ���� <[email protected]> wrote:
>> >> Dear Statalisters,
>> >>
>> >> Following the suggestion at
>> >http://www.stata.com/support/faqs/res/statalist.html#noa>nswer,
>> >
>> >> I am rephrasing the question to which I seem to have
>> >no response.
>> >>
>> >> I am programming a nested algorithm in which the 'outer'
>> >code is maximum likelihood, and the 'inner' one is the dynamic
>> >programming code I wrote in mata. I need to 1) take the
>> >coefficients obtained after first iteration of maximum
>> >likelihood (outer) algorithm, such as those reported by ml
>> >maximize, trace, 2) use them in my mata code to solve the
>> >dynamic programming problem inside mata, 3) take the results
>> >of that solution back to ml algorithm to obtain the estimates
>> >at the next iteration, and 4) continue this procedure until
>> >convergence.
>> >
>> >*
>> >* For searches and help try:
>> >* http://www.stata.com/support/faqs/res/findit.html
>> >* http://www.stata.com/support/statalist/faq
>> >* http://www.ats.ucla.edu/stat/stata/
>> >
>>
>> *
>> * For searches and help try:
>> * http://www.stata.com/support/faqs/res/findit.html
>> * http://www.stata.com/support/statalist/faq
>> * http://www.ats.ucla.edu/stat/stata/
>>
>
>*
>* For searches and help try:
>* http://www.stata.com/support/faqs/res/findit.html
>* http://www.stata.com/support/statalist/faq
>* http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
