A simplification to support vector machine for the second training

作者：Fang, Jinglong^*; Chen, Shuo; Pan, Zhigeng; Wang, Yigang

来源：16th International Conference on Artificial Reality and Telexistence, Hangzhou, PEOPLES R CHINA, 2006-11-29 To 2006-12-02.

For complicated recognition problem, the number of support vectors is large and recognition speed is low, because some sample were divided into section by error this time. To solve this problem, a method is bought to simplify the support vector machines based the minimal misestimate margin idea. Experiments show that this new support vector machine not only reduces the number of support vectors and recognition time but also has the same accuracy as (even better than) traditional support vector machine.

• 单位
  计算机辅助设计与图形学国家重点实验室; 杭州电子科技大学; 浙江大学