基于特权加权孪生支持向量机的多视图学习
Multi-view learning with privileged weighted twin support vector machine
摘要：
加权孪生支持向量机（WLTSVM）挖掘尽可能多的样本中潜在的相似性信息，以改善非平行平面分类器的共同缺点。与双支持向量机（TWSVM）相比，它通过类间K近邻（KNN）删除多余的约束，降低了时间复杂度。多视图学习是机器学习的一个新的发展方向，它主要研究从多个特征集合表示的数据中学习获取信息。在本文中，我们提出了特权加权双支持向量机（MPWTSVM）的多视图学习。它既继承了WLTSVM的优点，又有其自身的特点。首先，从同一视角挖掘类内信息，提高泛化能力。其次，利用类间信息减少了冗余约束，从而提高了运行速度。最重要的是，它可以同时遵循一致性和互补性原则作为一个多视图分类模型。一致性原则通过最小化原目标函数中两个视图的耦合项来实现。互补原则是通过建立特权信息范式和MVL来实现的。一个标准的二次规划求解器被用来解决这个问题。与SVM-2K、MVTSVM、MCPK和PSVM-2 V等多视图分类模型相比，该模型具有更好的分类精度和分类效率。在45个二值数据集上的实验结果证明了该方法的有效性。
英文摘要：
Weighted twin support vector machines (WLTSVM) mines as much potential similarity information in samples as possible to improve the common short-coming of non-parallel plane classifiers. Compared with twin support vector machines (TWSVM), it reduces the time complexity by deleting the superfluous constraints using the inter-class K-Nearest Neighbor (KNN). Multi-view learning (MVL) is a newly developing direction of machine learning, which focuses on learning acquiring information from the data indicated by multiple feature sets. In this paper, we propose multi-view learning with privileged weighted twin support vector machines (MPWTSVM). It not only inherits the advantages of WLTSVM but also has its characteristics. Firstly, it enhances generalization ability by mining intra-class information from the same perspective. Secondly, it reduces the redundancy constraints with the help of inter-class information, thus improving the running speed. Most importantly, it can follow both the consensus and the complementarity principle simultaneously as a multi-view classification model. The consensus principle is realized by minimizing the coupling items of the two views in the original objective function. The complementary principle is achieved by establishing privileged information paradigms and MVL. A standard quadratic programming solver is used to solve the problem. Compared with multi-view classification models such as SVM-2K, MVTSVM, MCPK, and PSVM-2V, our model has better accuracy and classification efficiency. Experimental results on 45 binary data sets prove the effectiveness of our method.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
