Non-symmetric Support Vector Machines

A novel approach to calculate the generalization error of the support vector machines and a new support vector machine–nonsymmatic support vector machine–is proposed here. Our results are based upon the extreme value theory and both the mean and variance of the generalization error are exactly ontained.

• Type: Conference
• Year: 2001
• Where: IWANN
• Authors: Jianfeng Feng
• Keywords: Generalization Error, IWANN 2001, Neural Networks, Support Vector, Support Vector Machines