一种基于N-Gram的垃圾邮件过滤方法研究  

ON APPROACH OF SPAM FILTERING BASED ON N-GRAM

作　　者：林伟[1,2] 柳荣其[2] 徐熙[2] 

机构地区：[1]四川警察学院计算机系,四川泸州646000 [2]西华大学数学与计算机学院,四川成都610039

出　　处：《计算机应用与软件》2010年第2期121-123,共3页Computer Applications and Software

基　　金：四川省青年软件创新工程项目(2007AA42)

摘　　要：为了能够有效提取邮件样本集的特征及提高垃圾邮件过滤系统的性能,介绍基于N-Gram的切分算法及语言模型,在其基础上,提出了一种改进的N-Gram切分算法,给出了一种结合N-Gram语言模型的贝叶斯过滤模型。实验结果表明,提出的方法有效地提高了垃圾邮件过滤的性能。In order to extract E-mail samples＇ features effectively and improve the performance of spam filtering system, this paper introduces segmentation algorithm and language model based on N-Gram. Then according to that an improved N-Gram segmentation algorithm is proposed,a Bayesian filtering model integrating the N-Gram model is given as well. Experimental results show that the improved approach is effective in improving the performance of spam filtering.

关 键 词：邮件过滤 N—Gram 贝叶斯模型 特征选择 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP311.56[自动化与计算机技术—计算机科学与技术]