NSC:一种新型的垃圾邮件过滤器

NSC:A Novel Spam Classification

借鉴生物免疫系统能有效的区别自体与非自体防御外部病源体入侵的工作机制,结合用户行为特征模式,提出了一种基于免疫机制的垃圾邮件过滤器模型NSC.在该模型中,给出了自体、非自体、抗体和抗原的定义与实现方式,建立了抗体的克隆选择、学习机制和生命周期模型.详细的描述了模型的训练与检测过程.试验结果表明,基于免疫机制的过滤器不仅有效的提高了检测率,而且还表现出良好的自学能力与自适应性.

Inspiring by the mechanism that the immune system can discriminate between self and non-self effectively to protect body from pathogeny intrusion, and combining the activities pattern of the client user, an immunity-based spam classification was proposed in this paper. Self, nonself, antibody and antigen are defined. Thus clone selection, learning scheme and lifeey- ele of antibody are built. The training and detection procedure are described in detail. The experimental result shows that the new model not only increases the recall and precision rates, but demonstrates that the model has the features of self-learning and self-adaption.