A Novel Multi-Class Cluster SVM for Handwritten Chinese Character Recognition

This paper proposes a novel multi-class cluster support vector machine, which borrows ideas of nonparallel hyperplanes from generalized eigenvalue support vector machines. For a k-class classification problem, it trains k nonparallel hyperplanes respectively, and each one lies as close as possible to self-class while apart from the rest classes as far as possible. Then, the label of a new sample is determined by the class of its nearest hyperplane belonging to. Finally, the proposed method is applied to tasks of financial handwritten Chinese character recognition task, and preliminary experimental results show that its testing accuracy outperforms traditional multi-class support vector machines methods, in both linear and nonlinear cases.