AdaCost: Misclassification Cost-Sensitive Boosting

Abstract

AdaCost, a variant of AdaBoost, is a misclassification cost-sensitive boosting method. It uses the cost of misclassifications to update the training distribution on successive boosting rounds. The purpose is to reduce the cumulative misclassification cost more than AdaBoost. We formally show that AdaCost reduces the upper bound of cumulative misclassification cost of the training set. Empirical evaluations have shown significant reduction in the cumulative misclassification cost over AdaBoost without consuming additional computing power. 1 Introduction Recently, there has been considerable interest in costsensitive learning [15, 8, 7, 17, 6, 2]. Turney [15, 16] discusses learning tasks sensitive to the costs of misclassification among others. We are interested in reducing misclassification cost. It can be either constant for each type of misclassification or conditional on a specific example under di#erent types of misclassification. In troubleshooting systems, for examp...