wilmott.com

any good references on log-likelihood estimation (even better if includes code building hints)? other than Aaron's post on another thread:QuoteOriginally posted by: AaronMaximum likelihood is a simple technique for fitting parameters. It has many bad properties, but it does give analytic solutions where other methods cannot.Usually it's easier to work with the log likelihood. Write down the probability density function of your observations using variables for the parameters. Take the log. Take the derivative with respect to the parameters and set it to zero. You're done.as well, any views on whether Matlab is suitable for doing that, as i've heard something in the lines of Gauss being the only tolerable place to do it (if using the inbuilt routines, algorithms (BFGS and what not))?

Yinya,Any basic (upper-level undergrad/masters) econometrics text book will have a good explanation of maximum likelihood, better than you are likely to get from an online forum.As for Matlab, I estimate maximum likelihood models in it all the time. It and Gauss are different, and it might be true that Gauss is better at ML than Matlab, but that is only because Gauss was originally written for economists & statistcians, whereas Matlab was originally written for scientists and engineers, so the functions available are somewhat different. If you want to estimate ML models, you will need either (a) the statistics & optimization toolboxes, or (b) the CompEcon toolbox from here or (c) James P. LeSage's Econometrics toolbox (Google for it).HTH,matt.

thanks!will go that route.k.

I agree...go to James LeSage's website, look for the function called "maxlik"...he also attached a demo function as well.