[Scipy-tickets] [SciPy] #1097: stats.anderson wrong results for gumbel?

SciPy Trac scipy-tickets@scipy....
Fri Jan 22 08:30:27 CST 2010


#1097: stats.anderson wrong results for gumbel?
-------------------------+--------------------------------------------------
 Reporter:  josefpktd    |       Owner:  somebody
     Type:  defect       |      Status:  new     
 Priority:  normal       |   Milestone:  0.8.0   
Component:  scipy.stats  |     Version:  0.7.0   
 Keywords:               |  
-------------------------+--------------------------------------------------

Comment(by josefpktd):

 I applied the change to the usage of fit instead of the moment estimator
 in r6210

 I tried several examples and they look good, but no tests added yet.

 Clarification to previous comment: the moment estimator is supposed to
 estimate loc and scale, but the results look closer to mean and standard
 deviation, which is different.

 Enhancement proposal

  * allow ad test with given parameters instead of estimating them
  * provide ad as a distance measure for any distribution, even if critical
 values are not available
  * in future bootstrap estimator for critical values would be useful
  * check different versions of ad statistic (sup versus integral ?)

 and of course verify and add more tests

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/1097#comment:2>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list