作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (04): 266-268. doi: 10.3969/j.issn.1000-3428.2012.04.087

• 开发研究与设计技术 • 上一篇    下一篇

一种适用于应用层协议的特征提取算法

张玉冲,蔡皖东,丁要军   

  1. (西北工业大学计算机学院,西安 710129)
  • 收稿日期:2011-07-28 出版日期:2012-02-20 发布日期:2012-02-20
  • 作者简介:张玉冲(1987-),男,硕士研究生,主研方向:网络与信息安全;蔡皖东,教授、博士生导师;丁要军,博士研究生
  • 基金资助:
    国家“863”计划基金资助项目(2009AA01Z424)

Feature Extraction Algorithm for Application Layer Protocol

ZHANG Yu-chong, CAI Wan-dong, DING Yao-jun   

  1. (College of Computer, Northwestern Polytechnical University, Xi’an 710129, China)
  • Received:2011-07-28 Online:2012-02-20 Published:2012-02-20

摘要: 基于PrefixSpan算法,提出一种适用于应用层协议的特征提取算法。通过加入关于位置的约束,减少频繁序列模式的产生数量,结合特征提取过程的实际情况加入约束条件,从而在挖掘过程中减少投影数据库产生的个数,在构建投影数据库过程中,去除关于非频繁项的存储及投影数据库中序列数小于最小支持度的扫描过程。实验结果表明,与原算法相比,该算法的运行时间较短,提取的特征具有较高的准确率和较低的误报率。

关键词: 序列模式挖掘, 特征提取, PrefixSpan算法, 关联规则

Abstract: This paper proposes a feature extraction algorithm for application layer protocol based on PrefixSpan algorithm. It adds the constraints about position to reduce the number of frequent sequence mode, and is combined with the practical situation of the feature extraction process to join constraint conditions. In the mining process, it reduces the number of projection database. In construction of the projection database, it removes the storage about the frequent items and the scanning of which sequence number is less than the minimum support degree in projection database. Experimental results show that compared with the original algorithm, the running time of this algorithm is shorter, and the extraction features have higher precision and lower false alarm rate.

Key words: sequential pattern mining, feature extraction, PrefixSpan algorithm, association rule

中图分类号: