You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Thomas Neidhart (JIRA)" <ji...@apache.org> on 2015/06/24 11:02:04 UTC

[jira] [Commented] (MATH-1239) Add a fast sequential random sampler method

    [ https://issues.apache.org/jira/browse/MATH-1239?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14599141#comment-14599141 ] 

Thomas Neidhart commented on MATH-1239:
---------------------------------------

Knuth's method S can also be considered: http://rosettacode.org/wiki/Knuth%27s_algorithm_S

> Add a fast sequential random sampler method
> -------------------------------------------
>
>                 Key: MATH-1239
>                 URL: https://issues.apache.org/jira/browse/MATH-1239
>             Project: Commons Math
>          Issue Type: Sub-task
>            Reporter: Thomas Neidhart
>
> Many applications need a fast way to randomly select n samples from a population without replacement.
> So far the best way seems to be the method proposed by Vitter in the following paper: 
> http://www.mathcs.emory.edu/~cheung/papers/StreamDB/RandomSampling/1984-Vitter-Faster-random-sampling.pdf
> A comparison with other methods can be found here (same author):
> https://hal.archives-ouvertes.fr/inria-00075929/document



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)