[jira] [Resolved] (IMPALA-7426) T-test is an unreliable method for comparing non-normal distributions

["Jim Apple (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/IMPALA-7426?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jim Apple resolved IMPALA-7426.
-------------------------------
       Resolution: Fixed
    Fix Version/s: Impala 3.1.0

https://github.com/apache/impala/commit/cb26d8d82880e934aae9ef211ca7752de510e84a

> T-test is an unreliable method for comparing non-normal distributions
> ---------------------------------------------------------------------
>
>                 Key: IMPALA-7426
>                 URL: https://issues.apache.org/jira/browse/IMPALA-7426
>             Project: IMPALA
>          Issue Type: Bug
>          Components: Infrastructure
>            Reporter: Jim Apple
>            Assignee: Jim Apple
>            Priority: Major
>             Fix For: Impala 3.1.0
>
>
> https://en.wikipedia.org/wiki/Student%27s_t-test is for normally distributed variables, and many Impala benchmarks won't be formally distributed. In particular, none of them will have negative run times.
> We should consider https://en.wikipedia.org/wiki/Nonparametric_statistics like https://en.wikipedia.org/wiki/Mann%E2%80%93Whitney_U_test#Comparison_to_Student's_t-test in report_benchmark_results.py.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)