基于神经网络模型的垃圾邮件过滤器的设计与实现

Design and Realization of Spam Filtering Model Based on Neural Model

该文介绍了神经网络模型在垃圾邮件过滤中的应用。首先对通过浏览器收集到的邮件进行分析,将其转换为HTML源代码的形式,再根据HTML语言的特点对其进行特征提取,从而达到邮件预处理的目的。随后又采用LVQ神经网络建立分类器模型,以达到最终分离正常邮件(ham)和垃圾邮件(spam)的目的,对比实验表明,结合HTML代码的特征提取和LVQ神经网络的分类器模型效果更好。

This paper introduce the neural network model,especially discusses the application of the neural network in spam filtering.Firstly,content of the e-mail are analysed,converting it into form of html source code.Then we do feature extraction according to html language characteristic,so as to achieve the purpose of e-mail pre-processing.Finally,the pre-processing e-mail are using LVQ neural networks to design a classifier model in order to realize the purpose of separate ham from spam.The comparision test result show that the LVQ neural network classifier based on html characteristic has chieved relatively good results.