bessel K function - Gretl-devel - gretlml.univpm.it

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

bessel K function

WG: debugger

WG: [Gretl-users] autoreg and...

Allin Cottrell

Thursday, 18 June 2009 Thu, 18 Jun '09

3:49 p.m.

On Wed, 17 Jun 2009, Allin Cottrell wrote:

On an initial scan it seems that most things are fine but the Bessel 'K' function goes a bit strange for large values of X (as you say, values < 20 are OK).

Now fixed, courtesy of the netlib/R code. Allin.

Reply

Show replies by date

Klein, Christoph

Friday, 19 June Fri, 19 Jun

4:05 a.m.

New subject: AW: bessel K function

thank you *so* much!!! this function is much more accurate than my version. Additionaly, I propose to dump my bessel_I function, too. With the attached patch bessel_I for negative orders is calculated using the new besselK function. bessel_I for negative orders can't be calculated with the cephes function, because cephes assumes symmetry around 0 for the order (see code), which is false. Christoph ________________________________________ Von: gretl-devel-bounces(a)lists.wfu.edu [gretl-devel-bounces(a)lists.wfu.edu] im Auftrag von Allin Cottrell [cottrell(a)wfu.edu] Gesendet: Donnerstag, 18. Juni 2009 21:49 An: Gretl development Betreff: [Gretl-devel] bessel K function On Wed, 17 Jun 2009, Allin Cottrell wrote:

On an initial scan it seems that most things are fine but the Bessel 'K' function goes a bit strange for large values of X (as you say, values < 20 are OK).

Now fixed, courtesy of the netlib/R code. Allin. _______________________________________________ Gretl-devel mailing list Gretl-devel(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-devel

Reply

Allin Cottrell

10:14 a.m.

New subject: AW: bessel K function

On Fri, 19 Jun 2009, Klein, Christoph wrote:

thank you *so* much!!! this function is much more accurate than my version.

I can't claim credit for the revised version; it comes from the netlib "specfun" repository.

Additionaly, I propose to dump my bessel_I function, too. With the attached patch bessel_I for negative orders is calculated using the new besselK function. bessel_I for negative orders can't be calculated with the cephes function, because cephes assumes symmetry around 0 for the order (see code), which is false.

Thanks. I've applied your patch. I've also added (minimal) docmentation for the bessel function. One little refinement: you don't need to quote the first argument. For example, x = bessel(J, 2, 5.6) will do OK. Allin.

Reply

5630

days inactive

5631

days old

gretl-devel@gretlml.univpm.it

Manage subscription

2 comments

2 participants

tags (0)

participants (2)

Allin Cottrell
Klein, Christoph