一种改进的内网用户行为审计模型研究  

A Research on an Improved Intranet Users' Behavior Audit Model

作　　者：乔佩利[1] 李明明[1] 

机构地区：[1]哈尔滨理工大学计算机科学与技术学院,黑龙江哈尔滨150080

出　　处：《哈尔滨理工大学学报》2011年第5期57-60,共4页Journal of Harbin University of Science and Technology

基　　金：哈尔滨市科技攻关项目(200812AA2CG037)

摘　　要：目前,安全审计系统多是基于广域网网络数据的审计,对于内网用户行为审计还处于研究的初级阶段.传统的用户行为审计模型智能性低,集中式结构审计负担重,针对这一缺陷,提出了一种分布式内网用户行为审计模型,采用三级Agent结构,并融入了改进的遗传算法(IGA)的神经网络分析,最终实现对用户行为数据的分布采集,过滤传输和智能审计.Currently,most of the security audit systems are based on the audit of wide area network（WAN） data,and the audit of intranet users＇ behavior is still in the primary stage of research.The traditional audit model of users＇ behavior has low intelligence and the centralized structure audit has heavy burden.To deal with this deficiency,this thesis proposes a distributed intranet users＇ behavior audit model.This model adopts three-level Agent structure and incorporates improved neural network analysis of IGA.Finally,the distributed data acquisition,filtration and transmission,and intelligent audit of users＇ behavior are realized.

关 键 词：内网安全 行为审计 AGENT 遗传算法 神经网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]