Re: distribution fitting

*To*: mathgroup at smc.vnet.net*Subject*: [mg90368] Re: distribution fitting*From*: Bill Rowe <readnews at sbcglobal.net>*Date*: Mon, 7 Jul 2008 05:08:31 -0400 (EDT)

On 7/6/08 at 7:20 AM, y.eaidgah at gmail.com (Youness Eaidgah) wrote: >Dear friends, I have a set of data and I want to fit a distribution >on this data.I know that some software like Bestfit do this and also >incorporate several tests like goodness-of-fit tests, c2 test, >Kolmogorov-Smirnov test, and Anderson-Darling test, and finally >parameter estimation. is it possible in Mathematica? All of the above can be done in Mathematica. Note, several of the statistics above are not built into Mathematica. So, you will have to create the code to compute the statistic. >can Mathematica find the most suitable distribution for a set of data? No. Mathematica can be set up to fit various distributions to a data set and select one or more based on any criteria you specify when writing your code. But while it is possible to do this, it is almost certainly not a useful thing to do. Using some test statistic such as Kolmogorov-Smirnov to select a "best" distribution is really a mis-use of the statistic. You really should be using other knowledge about the data to determine an appropriate distribution to use. >can it estimate the distribution parameters also? Yes. Details of how to do this depend on the distribution you are working with and what kind of estimate you want.