Imbalanced SVM learning with margin compensation

Abstract

The paper surveys the previous solutions and proposes further a new solution based on the cost-sensitive learning for solving the imbalanced dataset learning problem in the support vector machines. The general idea of cost-sensitive approach is to adopt an inverse proportional penalization scheme for dealing with the problem and forms a penalty regularized model. In the paper, additional margin compensation is further included to achieve a more accurate solution. As known, the margin plays an important role in drawing the decision boundary. It motivates the study to produce imbalanced margin between the classes which enables the decision boundary shift. The imbalanced margin is hence allowed to recompense the overwhelmed class as margin compensation. Incorporating with the penalty regularization, the margin compensation is capable to calibrate moderately the decision boundary and can be utilized to refine the bias boundary. The effect decreases the need of high penalty on the minority class and prevents the classification from the risk of overfitting. Experimental results show a promising potential in future applications. © 2008 Springer-Verlag Berlin Heidelberg.