Re: Re: fit a BinomialDistribution to exptl data?
- To: mathgroup at smc.vnet.net
- Subject: [mg80756] Re: [mg80727] Re: fit a BinomialDistribution to exptl data?
- From: DrMajorBob <drmajorbob at bigfoot.com>
- Date: Thu, 30 Aug 2007 23:51:26 -0400 (EDT)
- References: <11420600.1188458042685.JavaMail.root@m35>
- Reply-to: drmajorbob at bigfoot.com
> It is very instructive to see how the results vary from sample to sample. Yes, results vary wildly (for all the methods). In just a few trials of 200 draws from BinomialDistribution[20, 0.2], I got n anywhere from 13 to 96. In more extensive testing (300 trials), the simple moment estimator (setting theoretical mean and variance equal to the sample mean and variance, then solving for n and p) returned n from -1447 to +521, and the other methods were not much better. Darren Glosemeyer's FindFit solver failed to converge on one sample, returning n greater than 4000. And so forth. Statistics is more art than science. Bobby On Thu, 30 Aug 2007 01:34:27 -0500, P_ter <peter_van_summeren at yahoo.co.uk> wrote: > This fitting with Mathematica is quite impressive for me. But it is not > enough just to fit. > I did in R with the package gamlss {n,p=.2} experiment with one sample. > The results were nearly the same (correction for sample size). > I give some {n,p} results from Ray's program: > {3.88, 3.25186},{23.9666, 0.161892},{22, 0.176364} > {3.965, 2.99877},{16.2707, 0.24369},{16, 0.247813} > {3.775, 3.16018},{23.1784, 0.162867},{21, 0.179762} > It is very instructive to see how the results vary from sample to > sample. And this is for fitting! > It could be a very nice demonstration for findFit: to see the difference > between the theoretical distribution, the sample and the found > distribution. > Very impressive what one can do with a few lines of code! > with friendly greetings, > P_ter > > -- DrMajorBob at bigfoot.com