Twin support vector machine based on capped L1-norm

In view of the low generalization performance of twin support vector machine model which is easily affected by outliers, a new twin support vector machine model based on capped L1-norm is proposed in this paper. The capped L1-norm with upper bound value is used to construct the optimization problem instead of L2-norm, which weakens the influence of outliers and noise points on the construction of two hyperplanes to a certain extent and enhances the robustness of the model. In addition, a simple and efficient iterative algorithm is proposed for the construction of a new twin support vector machine model optimization problem, and the convergence of the algorithm is proved theoretically. Experimental results on noiseless and noisy UCI data sets show that the proposed model is more robust and stable than other SVM models.