Open erineidschun opened 4 years ago
So I believe the problem I am working on is slightly misspecified, and so I'll end up with a scalar likelihood function. In any case, I'll leave this up, if anyone is interested. From my research and how I think about likelihood functions intuitively, it makes sense for it to be scalar ONLY : )
I was wondering if it is possible to do MLE when the likelihood function is a vector, not a scalar, and if so, how to modify our regular crit, or log_lik functions to incorporate a vector. Will scipy.optimize still run if the log_lik function returns an array instead of a scalar? I can't really find any resources online (or perhaps am looking up wrong keywords).
Thanks!