[Scipy-tickets] [SciPy] #758: stats.rdist strange behavior

SciPy scipy-tickets@scipy....
Fri Oct 24 15:14:59 CDT 2008


#758: stats.rdist strange behavior
-------------------------+--------------------------------------------------
 Reporter:  travis       |        Owner:  somebody
     Type:  defect       |       Status:  new     
 Priority:  normal       |    Milestone:  0.7.0   
Component:  scipy.stats  |      Version:  devel   
 Severity:  normal       |   Resolution:          
 Keywords:               |  
-------------------------+--------------------------------------------------
Comment (by josefpktd):

 The source of the error, I think, is in special.hyp2f1 which is incorrect
 for some values, see Ticket #759

 stats.rdist.rvs has also errors as a consequence, e.g. samples that are
 out of bounds

 generic cdf calculation in stats.rdist.veccdf(x,c) seems to give correct
 answer,
 verified for some values with
 http://functions.wolfram.com/webMathematica/FunctionEvaluation.jsp?name=Hypergeometric2F1

 (However, I think there are still some other problems with the generic
 calculations)

 {{{
 x=0.75
 for c in [1,10,100, 150,200,300,400]:
     print 'c= ', c
     print 'hyp2f1 parameters: ', 0.5,1.0-c/2.0,1.5,x*x
     print 'hyp2f1  ', special.hyp2f1(0.5,1.0-c/2.0,1.5,x*x)
     print 'correct ',  (stats.rdist.veccdf(x,c) -
 0.5)*special.beta(0.5,c/2.0)/x
 }}}

 Josef

-- 
Ticket URL: <http://scipy.org/scipy/scipy/ticket/758#comment:1>
SciPy <http://www.scipy.org/>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list