The Extensions Of Nu-Support Vector Classification

Two extension models of nu-support vector classification (nu-SVC), the model called nu-SVC+ and another mixed model with noise, are investigated. They have the ability to learn the hidden information of training data which the conventional model is incapable. For the mixed model, when epsilon -> 1, the parameter nu has the significant that it is an upper bound on the fraction of margin errors and a lower bound on the fraction of support vectors, which is also testified by the experiments.