R help - Mar 2012 - Fitting & evaluating mixture of two Weibull distributions

Hello,

I would like to fit a mixture of two Weibull distributions to my data, estimate
the model parameters, and compare the fit of the model to that of a single
Weibull distribution.

I have used the mix() function in the 'mixdist' package to fit the mixed
distribution, and have got the parameter estimates, however, I have not been
able to get the log-likelihood for the fit of this model in order to compare it
to the single distribution.
I have also had a look at the 'Mixtools' and 'Flexmix' packages.
With both of these it is possible to fit a mixture of distributions, but, as far
as I can tell, these are pre-defined and do not have a function for a mixture of
two Weibull distributions.
I have used the 'fitdistrplus' package to fit the single distribution.

Does anyone have any suggestions either for getting the log-likelihood for
models fitted with the mix() function, or other packages / functions which allow
me to fit a double Weibull and get the log-likelihood for the model?

Any suggestions are much appreciated!

Best wishes,

Nienke


----------------------------------------------------------------------------------------------
Nienke Alberts
PhD Candidate
Centre of Research in Evolutionary & Environmental Anthropology
University of Roehampton


Consider the environment. Please don't print this e-mail unless you really
need to.

This email and any attachments are confidential and inte...{{dropped:21}}

The contribution of each observation to the logliklihood is
            log(p*f1(x) + (1-p)*f2(x))
where f1 and f2 are the two density functions.   Just sum.

  The standard likelihood ratio test is problematic here, as there
are two parametrizations that reduce the mixture model to a single
component: p=1 vs p < 1 plus the shape and scale parameters
of the two components are identical.

albyn

On 3/5/12 2:57 AM, N.Alberts at roehampton.ac.uk wrote:
> Hello,
>
> I would like to fit a mixture of two Weibull distributions to my
> data, estimate the model parameters, and compare the fit of the model
> to that of a single Weibull distribution.
>
> I have used the mix() function in the 'mixdist' package to fit the
> mixed distribution, and have got the parameter estimates, however, I
> have not been able to get the log-likelihood for the fit of this 
> model
> in order to compare it to the single distribution.
> I have also had a look at the 'Mixtools' and 'Flexmix'
packages. With
> both of these it is possible to fit a mixture of distributions, but,
> as far as I can tell, these are pre-defined and do not have a 
> function
> for a mixture of two Weibull distributions.
> I have used the 'fitdistrplus' package to fit the single 
> distribution.
>
> Does anyone have any suggestions either for getting the
> log-likelihood for models fitted with the mix() function, or other
> packages / functions which allow me to fit a double Weibull and get
> the log-likelihood for the model?
>
> Any suggestions are much appreciated!
>
> Best wishes,
>
> Nienke
>
>
> 
>
----------------------------------------------------------------------------------------------
> Nienke Alberts
> PhD Candidate
> Centre of Research in Evolutionary & Environmental Anthropology
> University of Roehampton
>
>
> Consider the environment. Please don't print this e-mail unless you
> really need to.
>
> This email and any attachments are confidential and 
> inte...{{dropped:21}}
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.