[
https://issues.apache.org/jira/browse/LUCENE-7305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15302161#comment-15302161
]
ASF subversion and git services commented on LUCENE-7305:
---------------------------------------------------------
Commit 55d854566e1e3c14cd91d91f414469104c935103 in lucene-solr's branch
refs/heads/branch_6x from [~teofili]
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=55d8545 ]
LUCENE-7305 - use macro average in confusion matrix metrics, removed unused
import in datasplitter
(cherry picked from commit dc50b79)
> Use macro average in confusion matrix metrics to normalize imbalanced classes
> -----------------------------------------------------------------------------
>
> Key: LUCENE-7305
> URL: https://issues.apache.org/jira/browse/LUCENE-7305
> Project: Lucene - Core
> Issue Type: Improvement
> Components: modules/classification
> Reporter: Tommaso Teofili
> Assignee: Tommaso Teofili
> Fix For: 6.1
>
>
> {{ConfusionMatrix}} multi class measures should be based on macro average to
> avoid bias (for the good or the bad) from imbalanced classes.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
