Robust AdaBoost based ensemble of one-class support vector machines

One-class support vector machine (OCSVM) is a commonly used one-class classification method for tackling novelty detection problems. Unfortunately, employing the traditional AdaBoost on it may not produce satisfying performance when there are outliers within training samples. In this paper, the conventional loss function of AdaBoost is redesigned, i.e., substituting the exponential loss function by a more robust one to enhance the robustness of the traditional AdaBoost based ensemble of OCSVMs. The proposed loss function is defined as the weighted combination of the modified exponential loss function and the squared loss function. The robust AdaBoost based on the proposed loss function is introduced by redesigning the update formulae for the weights of base classifiers and the probability distribution of training samples. The upper bounds of empirical error and generalization error for the robust AdaBoost based ensemble of OCSVMs are derived. Experimental results on the synthetic and benchmark data sets demonstrate that the proposed ensemble method is superior to its related approaches.