Search

In this paper, we propose a nonparallel support vector machine with pinball loss (Pin-NPSVM) that deals with the noise sensitivity and resampling instability of NPSVM. More specifically, we redefine a pinball loss funtion and build a pair of quantile hyper-planes. Each quantile hyper-plane is constructed by using the new pinball loss instead of ɛ-insensitive loss, which makes the new classification model be insensitive to noise samples, especially for feature noise samples around the decision boundary. Moreover, instead of hinge loss, Pin-NPSVM also builds a pair of decision boundaries based on traditional pinball loss, which further improves the anti-nosie ability of the classification model. In a word, Pin-NPSVM not only inherits the characteristics of the nonparallel optimal hyper-planes, but also has a consistent model with Pin-SVM, which can process noise data well. Finally, numerical experimental results show that the Pin-NPSVM has more obvious advantages than other models in classification performance, especially for noise datasets. [ABSTRACT FROM AUTHOR]

• Nonparallel support vector machine with large margin distribution is proposed. • The terms of margin mean and margin variance are introduced into the objective function. • The linear and nonlinear models have more powerful capability of generalization. The large margin distribution machine (LDM) combines the working principle of support vector machine (SVM) and the margin distribution to directly improve the algorithm's generalization. The margin distribution can be expressed with the margin mean and margin variance. It has been proved to be an efficient algorithm for binary classification. Inspired by the LDM, a novel classifier termed as LMD-NPSVM is proposed to improve the generalization performance of the nonparallel support vector machine (NPSVM) in this paper. Firstly, to meet the structure of NPSVM, the large margin distribution is reconstructed. Then, the linear LMD-NPSVM is built by introducing the reconstructed margin distribution into NPSVM. In addition, the linear case is extended to the nonlinear case with a kernel trick. All experiments show that our LMD-NPSVM is superior to the state-of-the-art algorithms in generalization performance. [ABSTRACT FROM AUTHOR]