设计了一种手机终端上基于短信内容的垃圾短信过滤系统。系统采用了平衡Winnow算法,该算法具有分类速度快、性能好以及支持在线更新的优点,适用于手机终端资源有限、需要实时或者定期更新分类器的情况。通过一系列的实验分析,证明该方法的有效性,并给出了对该方法的全面评估。对于该算法将来在信息过滤领域的应用,提供了全面的分析依据。
This paper proposed a content-based spam filter embedded in mobile phones. It used the balanced Winnow algorithm as the kernel classifier in the system, since the algorithm could classify texts faster, better and what' s more, it allowed on-line setting. The Winnow classifier was suitable for systems which had limited resources and needed to be renewed at inter- vals. Then implemented a series of experiments to prove the good performance of the system, and analyzed the method from many aspects. This paper gave a statistic support for the information filters using balanced Winnow.