摘要
由于信息系统中用户行为类型的多样性,采用单一的检测方法对内部合法用户的恶意行为进行检测评估效果不好。提出一种基于多实体贝叶斯网络的用户行为可信评估方法,利用多实体贝叶斯网络的对行为的多种要素进行建模、融合,共同推理出用户行为的可信程度。最后的仿真实验表明所提出的模型在信息系统中的用户行为可信评估具有较好的效果。
Since the type of user behaviour in information system is diverse, to use a single detection method to detect and evaluate malicious behaviours within internal legal users will not reach preferable effect. Thus we propose a MEBN-based trusted assessment method for user behaviour. The MEBN is employed for modelling and integrating various elements of user behaviour, then to infer the trusted degree of user behaviour. In the end of the paper, the simulation results show that the proposed model has good effect in trusted assessment of user behaviour in information system.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第12期214-218,共5页
Computer Applications and Software
关键词
用户行为
多实体贝叶斯网络
评估
User behaviour Multi-entity Bayesian network(MEBN) Assessment