基于行为和内容协作分析的垃圾短信过滤系统

Collaborative behaviours and contents analysis for SMS spam filtering system

随着移动网络的发展,垃圾短信已经成为亟需解决的重要课题。当前的垃圾短信过滤技术在应用于大规模垃圾短信过滤时,很难兼顾过滤的准确性和实时性要求。本文针对垃圾短信的特点,提出一种基于行为和内容协作分析的垃圾短信过滤机制并构建了原型系统。为了提升过滤及时性和过滤效果,采用行为分析为主、内容分析为辅的策略,并引入概率理论对归一化中文短信进行分析建模。该系统克服了当前垃圾短信过滤方法的不足,完善了垃圾短信过滤机制。通过真实垃圾短信数据对该原型系统进行验证,表明其可以大幅提高大规模垃圾短信的过滤效率和效果。

With the development of mobile networks, SMS spare has become an important challenge which need to be solved urgently. Current spam filtering technique is hard to fulfil the task accurately and timely in the presence of large-scale SMS systems. The paper proposes a mechanism based on collaborative behaviours and contents analysis and further develops a prototype system. In order to obtain a good filter result in time, we take the strategy of behaviour analysis, assisting with content analysis and introduce the probability theory into unitary short messages in Chinese. The proposed system improves the current SMS spam filtering methods by overcoming their disadvantages. By real SMS spam data running on the prototype system, we have showed that it can improve the effect of spam filter efficiently.