Home People Research Publications Demos
News Jobs Prospective
About Internal

Cost Sensitive SVM

A new procedure for learning cost-sensitive SVM classifiers is proposed. The SVM hinge loss is extended to the cost sensitive setting, and the cost-sensitive SVM is derived as the minimizer of the associated risk. The extension of the hinge loss draws on recent connections between risk minimization and probability elicitation. These connections are generalized to costsensitive classification, in a manner that guarantees consistency with the cost-sensitive Bayes risk, and associated Bayes decision rule. This ensures that optimal decision rules, under the new hinge loss, implement the Bayes-optimal costsensitive classification boundary. Minimization of the new hinge loss is shown to be a generalization of the classic SVMoptimization problem, and can be solved by identical procedures. The resulting algorithm avoids the shortcomings of previous approaches to cost-sensitive SVM design, and has superior experimental performance.

Experimental Results:
Fraud Detection and Business Decisions

The performance of the CS-SVM was evaluated on the German Credit data set. This dataset has 700 examples of good credit customers and 300 examples of bad credit customers. Each example is described by 24 attributes, and the goal is to identify bad costumers, to be denied credit. This data set is particularly interesting for cost-sensitive learning because it provides a cost matrix for the different types of errors. Classifying a good credit customer as bad (a false-positive) incurs a loss of 1. Classifying a bad credit customer as good (a miss) incurs a loss of 5. Using the CS-SVM algorithm results in a substantial reduction of cost by 37.36%.


Publications: Risk minimization, probability elicitation, and cost-sensitive SVMs
Hamed Masnadi-Shirazi and Nuno Vasconcelos.
International Conference on Machine Learning (ICML), 2010.
(acceptance rate 20%)

Contact: Nuno Vasconcelos, Hamed Masnadi-Shirazi


Copyright @ 2007 www.svcl.ucsd.edu