基于OWA算子与FSVM的邮件过滤

E-mail Filtering Based on OWA Operator and FSVM

不同用户对邮件的合法性有着不同的认识,因此对邮件过滤的研究应视为不确定信息处理问题,该文提出了一种融合有序加权平均(OWA)算子与模糊支持向量机(FSVM)的邮件过滤方法。其主要思想是利用OWA算子来产生每封邮件的一个合理的综合评价值作为其隶属度,采用FSVM对邮件进行分类。仿真实验结果验证了该方法的有效性。

Whether an e-mail is a legal one or not, different users hold different opinions. As a result, research of e-mail filtering should be considered as dealing with the uncertainties. This paper proposes an integrated method of ordered weighted averaging（OWA） operator and fuzzy support vector machines（FSVM）. Its main idea is to obtain a reasonable value of comprehensive evaluation as each mail＇s membership degree by using the OWA operator method, then classify each mail though FSVM. Simulative experiments are conducted to verify the effectiveness of the method.