摘要
针对电商平台中不可信用户的入侵行为问题,提出一种基于改进贝叶斯网络的用户行为认证方法。结合贝叶斯网络信念推理的特点和用户行为证据对用户行为等级进行认证;通过卡方检验(Chi-square test)筛选出在贝叶斯网络计算过程中占比重要的行为特征,依据特征与类别之间的相关性进行降维,提高认证精度;对认证模型进行分析得到用户行为证据集的权重,并将其应用到行为认证算法中,确保认证结果的安全性和可靠性。实验结果表明,该方法有效提高用户行为认证的准确率和执行效率,并为用户信息安全检测机制提供了新思路。
Aimed at the intrusion behavior of untrusted users in e-commerce platforms,a user behavior authentication method based on improved Bayesian networks is proposed.The characteristics of Bayesian network belief reasoning and user behavior evidence were combined to authenticate user behavior levels.Chi-square test was used to screen out the behavioral characteristics that were important in the Bayesian network calculation process.The dimensionality was reduced according to the correlation between characteristics and categories to improve authentication accuracy.The authentication model was analyzed to obtain the weight of the user behavior evidence set,and it was applied to the behavior authentication algorithm to ensure the safety and reliability of the authentication result.Experimental results show that this method effectively improves the accuracy and execution efficiency of user behavior authentication,and provides a new idea for user information security detection mechanism.
作者
陈楠
田立勤
毋泽南
朱洪根
Chen Nan;Tian Liqin;Wu Zenan;Zhu Honggen(School of Computer,North China Institute of Science and Technology,Beijing 101601,China;School of Computer,Qinghai Normal University,Xining 810000,Qinghai,China)
出处
《计算机应用与软件》
北大核心
2023年第8期330-336,共7页
Computer Applications and Software
基金
国家重点研发计划项目(2018YFC0808306)
河北省重点研发计划项目(19270318D)
河北省物联网监控工程技术研究中心项目(3142018055)
青海省物联网重点实验室项目(2017-ZJ-Y21)。
关键词
用户行为认证
行为证据集
卡方检验
贝叶斯网络
行为认证等级
User behavior authentication
Behavior evidence set
Chi-square test
Bayesian network
Behavior authentication level
