摘要
为提高入侵检测的精确性和有效性,通过对基本序列模式挖掘算法(Aprior算法)的分析,针对其缺点并结合入侵检测数据的特殊性,设计了改进的Aprior算法用于序列模式挖掘算法,算法将数据属性分成多个等级,侧重于多属性的序列模式挖掘,算法首先寻找高频轴属性值事件,再迭代降低支持度并增加新的低频轴属性值,用于比较长的频繁项集。同时以网络数据和日志文件数据为实验基础,从算法的精确性和适应性方面进行了比较。
An improved sequential patterns mining algorithm based on Aprior algorithm for intrusion detection is designed. It classifies data by the properties and focus on the sequential patterns by the multiple properties. It is effective for the long frequent item sets and improves the accuracy of intrusion detection largely. Its properties of accuracy and adaptation have been verified by analysis of audit data from network record and log files.
出处
《计算机技术与发展》
2008年第11期149-151,155,共4页
Computer Technology and Development
基金
广东省自然科学基金(04010589)
