Hi,thanks a lot for the answer. I think that my problem is before packages. I have a large amount of different datasets and I had a random look on their histograms. I know that some of them look like the combination I have explained (exponential+pareto, exponential+gamma) but I am not sure still.I wanted to know if there is a "preliminary" approach where a fitting algorithms tries some of the suggested combinations and give me hints where to focus my future research.I will have a look on the package you suggested. RegardsAlex
On Monday, December 8, 2014 11:23 PM, Spencer Graves <spencer.gra...@prodsyse.com> wrote: Have you considered "distr" and related packages? If this does not solve your problem, have you considered searching with "findFn" in the "sos" package? If that still does not produce sufficient enlightenment, please try this list again with "commented, minimal, self-contained, reproducible code" describing what you want in a bit more detail (as indicated at the end of emails on this list). Spencer On 12/8/2014 10:45 AM, Alaios via R-help wrote: > (I am sorry if you have received this email twice but it does not look sent > on my client) > > > > Hi all,I am having some heavy tailed data and I am trying to think of the >more appropriate package for the fitting.The canonical try should be something >like exponential and pareto or exponential + gamma (or gamma + gamma with >different shape parameters). I am trying to have one distribution that is >"classical" thin tailed one and another one which is more like a power law or >close (pareto,gamma,weibull, log-normal). > Do you have any recommendation with which packages I can start on and if the > functions you have in mind they can also provide information like AIC or BIC > so to help me choose the best combinations? > I would like to thank you in advance for your help > RegardsAlex > > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > 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. [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.