commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gilles Sadowski (Jira)" <j...@apache.org>
Subject [jira] [Commented] (MATH-1502) KolmogorovSmirnovTest#exactP value is incorrect in 3.6.1 potential bug
Date Wed, 30 Oct 2019 22:28:00 GMT

    [ https://issues.apache.org/jira/browse/MATH-1502?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16963481#comment-16963481
] 

Gilles Sadowski commented on MATH-1502:
---------------------------------------

The attached plot shows that the {{exactP}} function drops sharply around 0.55; this would
explain why alternative implementations produce either 0.6 or 0.1 for points around that value.

> KolmogorovSmirnovTest#exactP value is incorrect in 3.6.1 potential bug
> ----------------------------------------------------------------------
>
>                 Key: MATH-1502
>                 URL: https://issues.apache.org/jira/browse/MATH-1502
>             Project: Commons Math
>          Issue Type: Bug
>    Affects Versions: 3.6.1
>            Reporter: Ethan Mahintorabi
>            Priority: Critical
>         Attachments: exact_vs_approximate.png
>
>
> I believe that I have found a bug in the calculation of the exactP value in KolmogorovSmirnovTest.java
in version 3.6.1.
>  
> In version 3.4.1 KolmogorovSmirnovTest#exactP(0.5633587786259542, 3, 3, true) = 0.6
>  
> Where as
>  
> In version 3.6.1 KolmogorovSmirnovTest#exactP(0.5633587786259542, 3, 3, true) = 0.09999999999999998
>  
>  
> The approximateP value has stayed the same at ~=0.72. This seems to correspond with the
change in methods of calculation in e38bbb9f4191d0d21dea0ba31fdc131b97a5597b
>  
> [https://github.com/apache/commons-math/commit/e38bbb9f4191d0d21dea0ba31fdc131b97a5597b#diff-34100572fcc8e6537bc7445a0e82b900]
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Mime
View raw message