You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@commons.apache.org by "Rob Tompkins (JIRA)" <ji...@apache.org> on 2017/04/12 00:02:42 UTC

[jira] [Updated] (MATH-1235) Improve performance of DBSCAN clustering algorithm

     [ https://issues.apache.org/jira/browse/MATH-1235?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Rob Tompkins updated MATH-1235:
-------------------------------
    Fix Version/s: 4.X

> Improve performance of DBSCAN clustering algorithm
> --------------------------------------------------
>
>                 Key: MATH-1235
>                 URL: https://issues.apache.org/jira/browse/MATH-1235
>             Project: Commons Math
>          Issue Type: Improvement
>            Reporter: Thomas Neidhart
>             Fix For: 4.X
>
>
> The current implementation uses a naive proximity search which results in a quadratic runtime complexity. This could be improved by using a kd-tree for spatial indexing of the input data.
> Furthermore, the algorithm does not return the points considered to be noise.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)