基于神经网络的“垃圾”邮件过滤系统设计

Design of Spam Filtering System Based on Neural Network

为了对垃圾邮件进行有效地过滤,以神经网络作为分类器,采用由垃圾邮件发送者进行确认的邮件认证方法设计了邮件过滤系统。神经网络的自学习、自适应能力解决了垃圾邮件特征不断变化而过滤方法相对固定的矛盾。新的垃圾邮件认证方法使发送垃圾邮件比接收垃圾邮件更费时间,减少了用户收到垃圾邮件的数量。

In order to filter spams effectively,the spam filtering system is designed which adopts the e-mail identification method of identifying e-mail senders and has neural network as his categorizer. The self-studying and self-adapting ability of neural network resolves the contradiction that the features of spams varies constantly while the filtering methods are relatively static.As the new spam identification method is adopted, sending spams costs more time than receiving them,and the number of spams that user receives is reduced.