Abstract: Based on fuzzy of information contained in mail and asymmetry of legitimate mails and spam at the misjudgment price，a mail filtering method is proposed.It makes use of dual membership fuzzy support vector machine.According to provide a different pair of membership for each sample，the optimal classifier is derived.It improves the accuracy of mail filtering.The simulation results show that the method is able to effectively reduce the misjudgment of legitimate messages as spam.In additional，it has a high accuracy and so on.

Key words: spam filtering, fuzzy support vector machine, membership, dual membership fuzzy support vector machine

摘要： 针对邮件所含信息的模糊性和合法邮件与垃圾邮件错分代价的不对称性提出了基于双隶属度模糊支持向量机的邮件过滤方法，通过对每个样本赋予不同的双隶属度，得到最优分类器，提高了邮件过滤的正确率。经仿真实验证明，该方法能够有效降低将合法邮件误判为垃圾邮件，而且有很高的正确率等特点。

关键词: 垃圾邮件过滤, 模糊支持向量机, 隶属度, 双隶属度模糊支持向量机