RapidMiner

Precision Recall Curves and auPRC

Status: Released

Close to a necesity for evaluation of imbalanced binary classification problems.

6 Comments (6 New)
Comments

Hi,

This paper is interesting and covers the topic well : AUPRC

Good luck

Sven

RM Staff

Dear @SvenVanPoucke, Dear @yzan,

 

i've got a prototype opertor ready. It will hit operator toolbox as soon as i got time to write the documentation. if you need a preview version of it, please PM me.

 

Best,

Martin

RM Research

Hi @SvenVanPoucke, Hi @yzan,

 

Just for completness. The Operator Toolbox extension covers now since version 0.4.0 (Blog Post about 0.4.0 release) the AUPRC.

 

Best regards,
Fabian

Community Manager
Status: Released

Operator Toolbox Extension

Contributor II

It's great that we have the AUPRC value generated through the Operator Toolbox Extension. What would be much more useful is the Precision-Recall curves for a classifier (for any given threshold or cutoff value), especially when the dataset has a significant skew for the class labels. See the linked description about this, borrowed from the "Introduction to Data Mining" (2nd edition) by Tan et al. The intent is show the resultant PR-curve: PR-curve link (part 1)PR-curve link (part 2)

Community Manager

thanks @amitdeokar. It is my sneaking suspicion that this is being worked on as an improvement to the operator. Stay tuned.... cc @mschmitz