基于粗糙集与模糊支持向量机的模式分类方法研究

Method of Pattern Classification Based on FSVM and RS Theory

针对传统支持向量机(SVM)多分类算法分类效果欠佳的问题,研究基于粗糙集(RS)理论和模糊支持向量机(FSVM)多类算法的模式分类新方法。首先用RS属性约简方法去除冗余信息,然后用FSVM结合三叉分类树多类算法对约简后的样本分类。用本文方法在UCI数据库的数据集上做实验,与其他方法相比分类速度和精度显著提高,说明该方法是有效的。

Aiming at inefficiency problem of traditional support vector machines （SVM） multi-class algorithm,based on fuzzy support vector machine （FSVM） multi-class algorithm and rough set （RS） theory,a new method of pattern classification is researched. First,the RS method is used to eliminate redundant information with attribute reduction method. Then combining with triple decision tree algorithm,FSVM multi-class algorithm is used to classify the reduced samples. By Using the method proposed in this paper,simulation experiments are done on UCI database. Compared with other methods ,the results show speed and accuracy of classification were significantly increased. So the method is effective.