[SciPy-user] bug in modified Bessel function iv

Mark Bakker markbak@gmail....
Fri Sep 7 06:55:30 CDT 2007


Hello -

It seems that the modified Bessel function iv returns incorrect values for
large argument.

iv(0,100) gives 5.72185663838e+041

while the equivalent

jv(0,complex(0,100))
(1.07375170713e+042+0j )

This latter results is the correct answer, as verified in Abramowitz and
Stegun, Table 9.11, Page 428. Using the jv is a work-around, but this should
probably be fixed. The two results start to deviate for arguments over about
50, with jv giving the correct answer.

On a related note, there are implementations for Bessel functions of integer
order (jn, kn) but not for the modified Bessel function In. I guess this is
because the function would be called 'in', but it would be nice to have a
special function for integer order, and I am pretty sure they are around.

Thanks, Mark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/scipy-user/attachments/20070907/841d16b6/attachment-0001.html 


More information about the SciPy-user mailing list