[
https://issues.apache.org/jira/browse/MATH-1310?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Phil Steitz resolved MATH-1310.
-------------------------------
Resolution: Fixed
Fixed in dcd8015fa (master) and 2983ff81b (3_X).
The functions C and N defined by Wilcox have been implemented with recursive
definitions unwound. Using N, exactP has been reimplemented and is now fast
enough to extend the small sample range to cover up to 10000 sample product.
This allowed monteCarloP to be eliminated from use by the ksTest default
implementation. That function has been deprecated.
> Improve accuracy and performance of 2-sample Kolmogorov-Smirnov test
> --------------------------------------------------------------------
>
> Key: MATH-1310
> URL: https://issues.apache.org/jira/browse/MATH-1310
> Project: Commons Math
> Issue Type: Bug
> Affects Versions: 3.5
> Reporter: Phil Steitz
> Fix For: 3.6
>
>
> As of 3.5, the exactP method used to compute exact p-values for 2-sample
> Kolmogorov-Smirnov tests is very slow, as it is based on a naive
> implementation that enumarates all n-m partitions of the combined sample. As
> a result, its use is not recommended for problems where the product of the
> two sample sizes exceeds 100 and the kolmogorovSmirnovTest method uses it
> only for samples in this range. To handle sample size products between 100
> and 10000, where the asymptotic KS distribution can be used, this method
> currently uses Monte Carlo simulation. Convergence is poor for many problem
> instances, resulting in inaccurate results.
> To eliminate the need for the Monte Carlo simulation and increase the
> performance of exactP itself, a faster exactP implementation should be added.
> This can be implemented by unwinding the recursive functions defined in
> Chapter 5, table 5.2 in:
> Wilcox, Rand. 2012. Introduction to Robust Estimation and Hypothesis Testing,
> Chapter 5, 3rd Ed. Academic Press.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
