基于SVM的垃圾短信过滤系统  

Spam Messages Filtering System based on SVM

作　　者：李慧[1] 叶鸿[1] 潘学瑞[1] 段震[1] 张燕平[1] 

机构地区：[1]安徽大学计算机科学与技术学院,安徽合肥230601

出　　处：《计算机安全》2012年第6期34-38,共5页Network & Computer Security

基　　金：国家自然科学基金(61175046);安徽省自然科学基金(11040606M145);国家级大学生创新实践项目

摘　　要：传统的垃圾短信过滤方案,以垃圾短信中出现的敏感词作为判断的依据,却忽略了正常短信中出现的词对分类的贡献,并且由于短信用语的灵活性,特征提取难度较大。提出了一种基于svm算法对垃圾短信进行监控和过滤的方案,该方案根据短信内容、短信长度等特征,对短信文本进行向量空间的表示。通过机器学习的方式,对垃圾短信进行判断,过滤。相比传统方法而言,本系统在过滤准确度和效率两方面均获得大幅度提升。The traditional scheme to filter spam messages is generally based on the occurrence of sensitive words in them,but ignores the contribution made to the classification by words appearing in normal messages.And the flexibility of message phrases renders feature extraction rather difficult.This text presents a scheme to superintend and filter spam messages on the basis of SVM algorithm.This scheme expresses the text in way of vector space with reference to features like the content and length of messages.And it uses the method of machine learning to judge and filter spam messages.Compared with traditional schemes,this system performs better in terms of filtering efficiency and accuracy.

关 键 词：垃圾短信 短信过滤 机器学习 支持向量机 文本分类 特征提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]