Statistical distributions in Boost.Math
Recently I have been working on a few more Statistical Distributions for Boost.Math. These will be in release in 1.45 and some more in 1.46, (and soon, an update to http://sourceforge.net/projects/distexplorer/ Distribution Explorer, a mixed C++ and C# Windows application that allows you to explore the properties of various statistical distributions, & calculate the Cumulative Distribution Function CDF, Probability Distribution Function PDF, or quantiles/percentiles). inverse chisquare, inverse gamma distribution, inverse Gaussian, Wald. 1 I am looking for some examples of their use for the documentation. If you could to propose and/or help work up some not-too-complicated examples of applications of these (they seem to be popular with followers of the Reverend Bayes ;-) - please contact me privately. 2 Are there other distributions (or 'special' functions) that you, or your colleagues, friends or relations, would find useful? If you have suggestions or requests, please tell me. Thanks Paul --- Paul A. Bristow, Prizet Farmhouse, Kendal LA8 8AB UK +44 1539 561830 07714330204 pbristow@hetp.u-net.com
Paul A. Bristow wrote:
Recently I have been working on a few more Statistical Distributions for Boost.Math.
These will be in release in 1.45 and some more in 1.46,
(and soon, an update to http://sourceforge.net/projects/distexplorer/ Distribution Explorer, a mixed C++ and C# Windows application that allows you to explore the properties of various statistical distributions, & calculate the Cumulative Distribution Function CDF, Probability Distribution Function PDF, or quantiles/percentiles).
inverse chisquare, inverse gamma distribution, inverse Gaussian, Wald.
1 I am looking for some examples of their use for the documentation.
If you could to propose and/or help work up some not-too-complicated examples of applications of these (they seem to be popular with followers of the Reverend Bayes ;-) - please contact me privately.
2 Are there other distributions (or 'special' functions) that you, or your colleagues, friends or relations, would find useful?
If you have suggestions or requests, please tell me.
Thanks
Paul
I am needing cdf of doubly-noncentric F distribution, and having trouble finding code (execpt in mathematica). I have found a few papers, but haven't found time to review them.
-----Original Message----- From: boost-users-bounces@lists.boost.org [mailto:boost-users- bounces@lists.boost.org] On Behalf Of Neal Becker Sent: Friday, November 05, 2010 10:38 AM To: boost-users@lists.boost.org Subject: Re: [Boost-users] Statistical distributions in Boost.Math
Paul A. Bristow wrote:
Recently I have been working on a few more Statistical Distributions for Boost.Math.
If you have suggestions or requests, please tell me.
I am needing cdf of doubly-noncentric F distribution, and having trouble finding code (execpt in mathematica). I have found a few papers, but haven't found time to review them.
Hmm - looks non-trivial :-( http://mathworld.wolfram.com/NoncentralF-Distribution.html gives the complex details. As I recall, hypergeometric functions caused so much numerical trouble that they were withdrawn from TR1 (and not implemented in Boost.math)? Or is that something different? A quick Google references some promising looking articles, but I don't have access to most of them. If you can provide them as pdf, I'll think about it more carefully. Paul
Paul A. Bristow wrote:
-----Original Message----- From: boost-users-bounces@lists.boost.org [mailto:boost-users- bounces@lists.boost.org] On Behalf Of Neal Becker Sent: Friday, November 05, 2010 10:38 AM To: boost-users@lists.boost.org Subject: Re: [Boost-users] Statistical distributions in Boost.Math
Paul A. Bristow wrote:
Recently I have been working on a few more Statistical Distributions for Boost.Math.
If you have suggestions or requests, please tell me.
I am needing cdf of doubly-noncentric F distribution, and having trouble finding code (execpt in mathematica). I have found a few papers, but haven't found time to review them.
Hmm - looks non-trivial :-(
http://mathworld.wolfram.com/NoncentralF-Distribution.html gives the complex details.
As I recall, hypergeometric functions caused so much numerical trouble that they were withdrawn from TR1 (and not implemented in Boost.math)? Or is that something different?
A quick Google references some promising looking articles, but I don't have access to most of them. If you can provide them as pdf, I'll think about it more carefully.
Paul
I've found some references, but I'm no mathematician. The one I've toyed with, because it was easy to understand, is: http://www.ifpress.com/pdfs/comm22-897.pdf This is a numerical integration. This sounds interesting, but I haven't looked at it yet: http://onlinelibrary.wiley.com/doi/10.1111/j.1467-842X.1965.tb00036.x/abstra... Some others: http://onlinelibrary.wiley.com/doi/10.1111/j.1467-842X.1969.tb00102.x/abstra... http://fmwww.bc.edu/cef99/papers/Paolella.pdf
I am needing cdf of doubly-noncentric F distribution, and having trouble finding code (execpt in mathematica). I have found a few papers, but haven't found time to review them.
I'm sure that rang a bell somewhere, is what you're looking for not here already: http://www.boost.org/doc/libs/1_44_0/libs/math/doc/sf_and_dist/html/math_too... ? HTH, John.
John Maddock wrote:
I am needing cdf of doubly-noncentric F distribution, and having trouble finding code (execpt in mathematica). I have found a few papers, but haven't found time to review them.
I'm sure that rang a bell somewhere, is what you're looking for not here already:
http://www.boost.org/doc/libs/1_44_0/libs/math/doc/sf_and_dist/html/math_too...
?
HTH, John.
No, boost has singly-noncentral F distribution. I mean: http://mathworld.wolfram.com/NoncentralF-Distribution.html
No, boost has singly-noncentral F distribution. I mean:
That was the page I was looking at as well, but reading more carefully I see that while the give the PDF for the doubly non-central case, all the rest (including the CDF and the mathematica version referenced) are for the singly non-central. My mistake, John.
On Fri, Nov 5, 2010 at 4:59 AM, Paul A. Bristow <pbristow@hetp.u-net.com> wrote:
Recently I have been working on a few more Statistical Distributions for Boost.Math.
2 Are there other distributions (or 'special' functions) that you, or your colleagues, friends or relations, would find useful?
I recently needed to do some analysis of existing games and had to write myself a distribution for the sum of n dice with k sides. Its not a particularly difficult one to write, but many folks in the gaming industry wouldn't have a clue where to start, so it might be worthwhile putting in. -- Stirling Westrup Programmer, Entrepreneur. https://www.linkedin.com/e/fpf/77228 http://www.linkedin.com/in/swestrup http://technaut.livejournal.com
participants (4)
-
John Maddock -
Neal Becker -
Paul A. Bristow -
Stirling Westrup