最小风险的Naive Bayes技术在反垃圾邮件系统中的应用  

Risk Minimization Naive Bayes Tech Applying in Anti-spam Filtering

作　　者：张健[1] 陈拓[2] 韩益亮[1] 畅雄杰[1] 李彩霞[1] 潘峰[1] 

机构地区：[1]武警工程学院 [2]中国科学院西安分院,陕西西安710061

出　　处：《微电子学与计算机》2005年第12期139-141,共3页Microelectronics & Computer

摘　　要：文章提出了一套更好地过滤中文垃圾邮件的方案,这套方案将利用垃圾邮件规律的规则过滤和最小风险的NaiveBayes内容过滤算法结合了起来,并根据垃圾邮件的特性做了必要的改进。并且这套方案也在Linux/So-laris系统平台下基本上完成了大部分功能的软件编程。实际电子邮件服务器上对本方案进行了测试,结果显示这套方案取得了很好的过滤效果。最小风险的NaiveBayes技术是目前最重要的反垃圾邮件技术之一。Aim for the sake of filtering the Chinese spam effectively. Methods Rule-based filtering and risk minimization Naive Bayes are combined in this project, and this project is improved in some necessary places. The software of this project is completed mostly under the Linux/Solaris system platform. Results And this software system is tested on the Internet mail server. Test result shown, this project has a better effect of anti-spam filtering. Conclusion Risk minimization Naive Bayes tech is one of the most important anti-spam tech.

关 键 词：垃圾邮件 规则过滤 NAIVE BAYES 

分 类 号：TP311.56[自动化与计算机技术—计算机软件与理论]