Maximum Likelihood Estimation
oscar.j.benjamin at gmail.com
Sat Feb 2 22:41:12 CET 2013
On 2 February 2013 18:26, Wolfgang Keller <feliphil at gmx.net> wrote:
>> I am looking for a Python implementation of Maximum Likelihood
>> Estimation. If any one can kindly suggest. With a google search it
>> seems scipy,numpy,statsmodels have modules, but as I am not finding
>> proper example workouts I am failing to use them.
> For statistics I would suggest using R (http://www.r-project.org/)
> through RPy (http://rpy.sourceforge.net/).
> Both have dedicated mailinglists as well as extensive documentation.
I agree with Wolfgang that R is likely to be able to do what you want
and that you may have better luck asking this kind of question on
their mailing lists (or on the scipy mailing list).
In any case, though, you will need to be more specific about what you
mean. Maximum Likelihood Estimation (MLE) is a sufficiently vague
topic that there cannot really be an "implementation" of it. What kind
of model/data are you working with? Or are you working with pure
probability distributions? What kind of parameters are you trying to
find? Are the parameters you are trying to choose discrete or
continuous? Are you trying to find one parameter or several
simultaneously? Are you able to find an analytic solution that
transforms your MLE problem into a specific kind of mathematical
problem, such as solving a system of linear equations?
Assuming that you are able to compute directly the likelihood (or
log-likelihood) of whatever it is you are interested in, then your MLE
problem is simply an optimisation problem, so you may have better luck
searching for implementations of optimisation (you will still need to
answer the questions above to be able choose an optimisation method).
More information about the Python-list