Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: How to maximize an approximated likelihood

From   Feddag Mohand-Larbi <>
Subject   st: How to maximize an approximated likelihood
Date   Mon, 08 Apr 2013 10:04:56 +0200

Dear Statausers,

I am wondering to know how to maximize an approximated likelihood
depending on three parameters: the difficulty parameter as matrix, the
mean of the latent traits and its variance.

The function wrotten in Stata where the approximation is obtained by
Gauss-Hermite quadrature is given as follows:

qui gen proba=.
qui gen x=.
forvalues i=1/`nbobs' {
    local int2=1
    local u=RaschDepLoc[`i',1]
    local v=`diff'[1,1]
    local int1 "exp(`u'*(x-`v'))/(1+exp(x-`v'))"
    forvalues j=2/`nbitems' {
         local t=`diff'[`j',1]
         local r=RaschDepLoc[`i',`=`j-1'']
         local s=RaschDepLoc[`i',`j']
         local int2 
        local int="`int1'*`int2'"
        qui gausshermite `int', mu(`mu') sigma(`sigma') display
        qui replace proba=r(int) in `i'
gen lik=log(proba)*`nbobs'

where nbobs is the number of individuals, nbitems is the number of
items, RaschDepLoc is the matrix of data of order (nbobs*nbitems), diff
is the vector of difficulties parameters, mu and sigma^2 are the mean
and the variance of the latent traits.

I don't know how to write this function in order to use ml model lf myfunction (? =?)
and ml maximize


Dr Feddag

*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index