Kernel-Free Nonlinear Support Vector Machines for Multiview Binary Classification Problems

Multiview learning (MVL) frequently uses support vector machine- (SVM-) based models, but it can be difficult to select appropriate kernel functions and corresponding parameters. Then, by introducing kernel-free tricks, two multiview classifiers are proposed, called C -multiview nonlinear machine ( id="M2"> -MKNSVM) its id="M3"> ν -version, namely, id="M4"> -MKNSVM. They try find a quadratic hypersurface under each view classify the sample points employ consistency constraint fuse from views. Both primal dual problems of id="M5"> -MKNSVM id="M6"> do not involve functions; thus, they allowed solved directly. In addition, relationship solutions between is discussed in classifier. For id="M7"> -version id="M8"> MKNSVM, meanings their parameters them analyzed detail. The experimental results artificial benchmark datasets show that our methods superior some traditional MVL like SVM-2K, PSVM-2V, MvTSVM, especially id="M9">