tags:

views:

72

answers:

3

I have to estimate parameters of a mixture distribution which consists of Pareto distribution and Exponential distribution. I am using maximum likelihood estimation procedure by using log likelihood function and differentiating it with respect to each parameter.Since the log likelihood equation is non linear, I have to use Newton Rhapson iterative method.How can I do this using R or any other programming language?

+1  A: 

Have a look at the flexmix-package

Matti Pastell
A: 

ROOT has fitting facilities that make this relatively easy.

dmckee
Sure does but not that the question was not tagged C++.
Dirk Eddelbuettel
I believe C++ falls into the set of "any other programming language" but outside of the set of "languages which were tagged"
JD Long
@Dirk: Sure---but as JD says---I took the "other" languages bit as license to stray. However, it *is* worth mentioning that people who find R natural and comfortable may find ROOT unnatural and uncomfortable: the tools are specialized for different domains. Certainly, I won't be hurt if laks takes a different approach here...
dmckee
A: 

does optim not work?

alex