Information matrix for MLE

Discussion on Quantitative Methods in Finance
Forum rules
DISCLAIMER: We do not warrant or represent that this forum or its content is free of viruses, worms or other code that might be contaminating or destructive. We cannot guarantee that documents or files downloaded from the Site will be free from viruses and we do not accept any responsibility for any damage or loss caused by any virus. Accordingly, for your own protection, you must use virus-checking software when using the forum. You must not post or provide to us via the forum, any document or file which you believe may contain a virus. You must virus check any document or file which you intend to post or provide to us via the forum. You must ensure that any document or file you intend to post to the forum does not contravene any applicable laws or contravene any person's legal rights. We do not accept any responsibility for any damage or loss you may suffer.
giancarp
Posts: 3
Joined: Wed Mar 12, 2014 6:05 pm

Information matrix for MLE

Postby giancarp » Mon Apr 07, 2014 8:02 pm

I am computing the information matrix computing the Hessian of the loglikelihood function at the maximum likelihood estimates numerically.

In most cases I end up with meaningful information matrices, but when I have a small sample, the loglikelihood function is quite unstable and the second derivative computed numerically isn't usable.

My attention was caught by 203: "But sometimes even the second derivatives may be impossible to compute, in which case we can use an alternative form for the information matrix that requires only first derivatives of the log likelihood: I(theta) =E[score x score']"

Where can I find references about this? I am a bit confused since I am using my MLE and being a maximum I would expect that all the first derivatives should be zeros (and all the second derivatives should be negative - if one is positive then I've a saddle point). If all the first derivatives are zeros, the information matrix should be the same (and in fact this is what happens numerically).

What am I missing?

thank you very much for any suggestion or references,
gc

Return to “Volume I”

Who is online

Users browsing this forum: No registered users and 1 guest