Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: question concerning translation from Matlab to Mata
From
"Brian P. Poi" <[email protected]>
To
[email protected]
Subject
Re: st: question concerning translation from Matlab to Mata
Date
Thu, 10 Mar 2011 18:09:53 -0500
On 3/10/2011 5:55 PM, DC wrote:
Hi All,
I'm attempting to translate some code for an optimal stopping problem
that I had previously written in Matlab to Mata and I've encountered a
problem
in writing down a likelihood function.
A simplified version in Matlab would be a multinomial logit with three
choices, choice 0 parameter normalized to zero, :
...
[n k]=size(x);
D = 1+exp(x*b(1:k))+exp(x*b(k+1:2*k));
p_0 = 1./D;
p_1 = exp(x*b(1:k))./D;
p_2 = exp(x*b(k+1:2*k))./D;
lnlike=-sum( (y==0).*log(p_0) + (y==1).*log(p_1) +(y==2).*log(p_2));
....
However is it not clear how to specify and estimate something like
this using Mata
I tried the following:
void lmle(todo,b,y, x,lnf, S, H)
{
k = cols(x)
b1 = b[1::k,1]
b2 = b[k+1::2*k,1]
xb1 = x*b1'
xb2 = x*b2'
d = 1 :+ (exp(xb1) + exp(xb1))
p_0 = 1 :/ d
p_1 = exp(xb1) :/ d
p_2 = exp(xb2) :/ d
lnf = (y==0) :* log(p_0) + (y==1) :* log(p_1) + (y==2) :* log(p_2)
}
S = optimize_init()
optimize_init_evaluator(S,&lmle())
optimize_init_evaluatortype(S, "gf0")
optimize_init_argument(S,1,y)
optimize_init_argument(S,2,x)
optimize_init_params(S, J(1,cols(x),0))
b = optimize(S)
One immediate problem is that -optimize- passes the parameters as a row
vector, not a column vector, so the lines
> b1 = b[1::k,1]
> b2 = b[k+1::2*k,1]
should probably be
b1 = b[1, 1::k]
b2 = b[1, k+1::2*k]
That may or may not be the only problem.
-- Brian Poi
-- [email protected]
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/