Optimizing F-Measure with Support Vector Machines

Abstract

Support vector machines (SVMs) are regularly used for clas- sification of unbalanced data by weighting more heavily the error contribution from the rare class. This heuristic tech- nique is often used to learn classifiers with high F-measure, although this particular application of SVMs has not been rig- orously examined. We provide significant and new theoreti- cal results that support this popular heuristic. Specifically, we demonstrate that with the right parameter settings SVMs ap- proximately optimize F-measure in the same way that SVMs have already been known to approximately optimize accu- racy. This finding has a number of theoretical and practical implications for using SVMs in F-measure optimization.