摘要
鉴于传统的访问控制机制为用户提供静态的授权,极少监管用户行为,导致企业很难发现来自恶意的内部用户或账号失窃带来的数据安全威胁,定义和形式化描述了私有云下的用户行为,基于Hadoop架构提出了一个基于TensorFlow机器学习的用户行为分析的框架,通过用户行为数据采集、存储、特征选择、预处理,给出了建立并训练用于用户行为分析的机器学习模型的方法和过程,实现了企业私有云用户行为的自动分析,用于数据安全威胁的发现和响应.
Since traditional access control mechanisms provide static authorization for users,but with little supervision over their behaviors,it is difficult for enterprises to locate threats of data security posed by malicious insiders or compromised user accounts with high privilege. This article defines and formalizes user behaviors under Private Cloud,Through user behavior data acquisition,storage,feature selection and preprocessing,A Neural Network model for User Behavior Analysis is proposed with its training and optimization procedure. Thus a User Behavior Analysis model based on Tensor Flow Machine Learning platform over Hadoop framework is given,which can analysis user behaviors automatically,and can help enterprises to locate and respond to threats of data security efficiently in private cloud.
出处
《中南民族大学学报(自然科学版)》
CAS
北大核心
2017年第3期95-100,共6页
Journal of South-Central University for Nationalities:Natural Science Edition
基金
国家自然科学基金资助项目(61472121)
湖北省创新群体项目(2016CFA021)
关键词
私有云
数据安全
用户行为
机器学习
Private Cloud
Data Security
User Behavior
Machine Learning