[ 
https://issues.apache.org/jira/browse/SOLR-11160?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Joel Bernstein updated SOLR-11160:
----------------------------------
    Attachment: SOLR-11160.patch

> Add normalDistribution, uniformDistribution, sample and kolmogorovSmirnov 
> Stream Evaluators
> -------------------------------------------------------------------------------------------
>
>                 Key: SOLR-11160
>                 URL: https://issues.apache.org/jira/browse/SOLR-11160
>             Project: Solr
>          Issue Type: New Feature
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Joel Bernstein
>            Assignee: Joel Bernstein
>             Fix For: master (8.0), 7.1
>
>         Attachments: SOLR-11160.patch, SOLR-11160.patch, SOLR-11160.patch
>
>
> This ticket digs deeper into *probability distributions*, which are a key 
> foundation of statistics.
> Three new functions are introduced:
> 1) *normalDistribution*: This is a distribution function which returns a 
> normal distribution generator. It is designed to be used with both the 
> *sample* and the *kolmogorovSmirnov* functions. Eventually a function will be 
> added for every distribution supported by *Apache Commons Math*.
> 2) *sample*: Returns array of numbers that conforms to a given distribution. 
> You pass it a distribution function and size  and it returns the array that 
> conforms to the distribution.
> 3) *kolmogorovSmirnov*: Performs a goodness of fit test for distributions. 
> This function can will do both a one and two side test. The one sided test 
> compares a sample to a distribution function. The two sided test compares two 
> samples to see if that were pulled from the same distribution.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to