摘要
针对互联网邮件中垃圾邮件占比暴增的问题,提出了一种基于分类代理MCP的动态检测算法.该方法基于近半年时间对校园网邮件宿主机及各代理虚拟机间传输的会话日志的采集,针对记录中各类投递状态及状态消息集进行了行为分析,最终达到对垃圾邮件的有效检测,从而为分拣提供依据.实验结果表明,在持续进行了若干频次的分类策略调节后,该检测算法的准确度可高达96.1%.该设计可对垃圾邮件宿主机及代理虚拟机的行为进行有效检测,从而彻底抑制垃圾邮件的产生.
In order to solve the problem of increasing the proportion of spam in Internet mail, a dynamic detection algorithm based on MCP is proposed. Based on the collection of the session logs collected from the campus network mail hosts and virtual agents in the past six months,the method analyzes all kinds of delivery status and status message set in the record, and achieves the result of effective spam detection finally, so as to provide the basis for sorting. The experi-mental results show that after a certain number of frequency classification strategy is adjusted,the highest accuracy of the detection algorithm is up to 96.1%. The design detects the behavior of spam host and virtual machine effectively, and completely suppresses the generation of spam.
出处
《南京师范大学学报（工程技术版）》
CAS
2017年第3期80-86,共7页
Journal of Nanjing Normal University(Engineering and Technology Edition)
基金
中国高等教育学会教育信息化专项课题(2016XXYB02)
