It sounds like a reasonable addition. Can you fine a JIRA?
On Sun, May 18, 2014 at 3:19 AM, Karol Grzegorczyk <[email protected]> wrote: > Hi, > > To get more statistics on classification results, in our project we > slightly extended ResultAnalyzer and ConfusionMatrix classes, adding > calculation of weighted average of precision, recall and F1 score. Do you > think that it would be useful to apply this modification to the trunk? We > can provide path (including unit tests). > > Regards, > Karol >
