基于粗集理论和SVM算法的模式分类方法

A METHOD OF PATTERN CLASSIFICATION ON LINE BASED ON ROUGH SET AND SVM ALGORITHM

本文提出一种将粗集方法与SVM算法结合起来的模式分类方法。利用粗集理论在处理大数据量、消除冗余信息等方面优势,减少SVM训练数据,克服SVM算法因为数据量太大,处理速度慢等缺点;同时,借助SVM良好的分类性能,对粗集约简后的最小属性子集进行分类,实现模式分类算法的快速性能、高识别率和抗干扰性强等优点。本文以手写体汉字的识别为例,说明本算法的实用性。

In this paper, we demonstrate a classification method which combines Rough Set and SVM Algorithm. In virtue of the ability Rough Set has to decease the amount of data and get rid of redundancy, the method can reduce amount of training data used and overcome SVM' s defect of slow running speed when process large data set. At the same time, by the aid of SVM algorithm the method can classify the core of property set so as to have extensiveness and high identification rate, and avoid disturbance. The recognition of handwriting Chinese is given as an example to show that the method can be used practically.