基于多分类器融合的垃圾短信处理系统

Anti-spam SMS system based on multi classifier fusion

本文提出了利用朴素贝叶斯、社交网络SNS和广告主挖掘多个分类器融合的垃圾短信处理方式,系统基于Hadoop处理平台,从用户发送内容和发送行为两个维度着手堆垃圾短信识别,可以大幅度提升目前垃圾短信治理的召回率和准确率。

In this paper, we propose use naive bayesian, social networks and extract advertisers to tap the multi classifier fusion Anti-spam system, the system based on Hadoop platform, according to SMS content and user's behavior to identify spam messages, then the classification accuracy will be improved greatly.