Hi everyone,
I am trying to run the LDA model from the Stan user guide which can be found here:
However, I think the likelihood calculation in the code is not right:
for (n in 1:N) {
real gamma[K];
for (k in 1:K)
gamma[k] = log(theta[doc[n], k]) + log(phi[k, w[n]]);
target += log_sum_exp(gamma); // likelihood;
}
In the formula, it also includes the prior probabilities (for \theta and \phi); however, in the code while calculating the target , likelihood, they are not included. I would appreciate your help on that whether there is a reason they are not included.