Application of Decision Tree SVMs Based on Class Distribution to Mental Tasks Recognition

摘要：

A new multi-class support vector machine combined with decision tree and SVM is proposed. The decision direction is determined by separability measure based on class distribution, which can reduce the influence of error accumulation. A data set of BCI competition 2005 is analyzed. The satisfactory results are obtained with the highest classification accuracy 80.8％.

作者: HUANG Ling ZHANG Aihua

作者单位: LANZHOU Univ. of Tech.Lanzhou,730050, China

会议类型: 国际会议

会议名称: 第七届国际测试技术研讨会

会议地点: 北京

会议语种:英文

在线出版日期: 2007-08-05（万方平台首次上网日期，不代表论文的发表时间）