Robust truncated L2-norm twin support vector machine

This paper proposes a new robust truncated L-2-norm twin support vector machine ((TSVM)-S-2), where the truncated L-2-norm is used to measure the empirical risk to make the classifiers more robust when encountering lots of outliers. Meanwhile, chance constraints are also employed to specify false positive and false negative error rates. (TSVM)-S-2 considers a pair of chance constrained nonconvex nonsmooth problems. To solve these difficult problems, we propose an efficient iterative method for (TSVM)-S-2 based on difference of convex functions (DC) programs and DC Algorithms (DCA). Experiments on benchmark data sets and artificial data sets demonstrate the significant virtues of (TSVM)-S-2 in terms of robustness and generalization performance.