一种改进的PrefixSpan应用层特征自动提取算法  

An Improved Prefix Span Application Layer Feature Auto-Extraction Algorithm

作　　者：彭大芹[1] 杨彩敏 黄德玲[1] 

机构地区：[1]重庆邮电大学,重庆400065

出　　处：《信息通信》2017年第1期18-20,共3页Information & Communications

基　　金：国家科技重大专项资助项目(No.2012ZX03001012)

摘　　要：针对基于DPI的网络流量识别中的指纹特征提取困难问题,文章提出了改进的基于Prefix Span算法的连续序列模式挖掘算法来提取应用层指纹特征。首先,改进的算法在挖掘应用层特征时加入了属性约束。其次,在进行频繁序列投影前,又引入剪枝策略。最后,通过实验表明该改进的算法在挖掘应用层指纹特征方面大大提高了效率,提取出来的特征可有效地应用于网络流量识别技术中。In this paper, we propose an improved continuous sequential pattern mining algorithm based on Prefix Span to extract the fingerprint characteristics of application layer for the problem of fingerprint feature extraction in network traffic identification based on DPI. First, the improved algorithm adds attribute constraints to mine application layer features. Secondly, pruning strategy is introduced before frequent sequence projection. Finally, the experimental results show that the improved algorithm can greatly improve the efficiency of mining the fingerprint characteristics of the application layer, and the extracted features can be effectively applied to the network traffic identification technology.

关 键 词：网络流量识别 数据挖掘 指纹特征 PREFIX Span算法 频繁序列 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]