基于贝叶斯网络和用户行为日志挖掘的行为信任研究

深入研究信任管理和行为信任的模型及方法，设计基于贝叶斯网络的信任预测和控制算法，综合利用聚类和分布密度函数设置算法参数，建立可量化的证据与信任等级之间的对应关系，算法可预测多属性下的行为信任等级.深入IIS和.Net底层实现可配置的信任管理插件，形成用户行为日志，为预测和控制算法提供证据, 免除了一般Web日志的清洗工作.实验数据表明算法的应用提高了服务器各项性能，并约束了用户的商业行为.

Study on Behaviour Trust Based on Bayesian Network and User Behavior Log Mining

Study on behavior trust and is prospective. This paper studies trust management and behavior trust, designs trust forecast and control Algorithm based on Bayesian Networks, which can predict trust grade under the multi-trust-attribute conditions, and Clustering Arithmetic and Distribution Density Function are used synthetically to set parameters and the corresponding relationship is built between quantitive evidence and trust grade. The configurable plug-in is implemented through going deeply into IIS and .NET Framework to create the user behavior log to provide evidences for forecast and control Algorithm, which avoids data cleaning. The pratical data have shown that multiple performance are enhanced and the trade behaviors of user are restricted.