BookmarkSubscribeRSS Feed

Using the ROC Convex Hull Algorithm to Select a Classifier

Started ‎12-06-2021 by
Modified ‎12-07-2021 by
Views 3,227

Rick Wicklin (@Rick_SAS)  contributed an article in The Do Loop (https://blogs.sas.com/content/iml/2021/11/15/2d-convex-hull-sas.html) describing two-dimensional complex hulls in SAS, specifically centered on the SAS/IML CVXHULL function.

 

A while ago, I created an algorithm based on Provost and Foster's work using a convex hull representation of several classifiers to select the optimal classifier based on cost-weighted classification accuracy. I presented my work at the Interface 2003 conference under the title "Cost-Sensitive Classifier Selection Using the ROC Convex Hull Method" and wish to share it with the SAS community. The paper and the SAS macro are included (attached). I hope that they will be of interest to some SAS community members.

Version history
Last update:
‎12-07-2021 07:27 AM
Updated by:

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

Register now!

Free course: Data Literacy Essentials

Data Literacy is for all, even absolute beginners. Jump on board with this free e-learning  and boost your career prospects.

Get Started

Article Tags