基于机器学习建模的口令脆弱性检测  

Password Vulnerability Detection based on Machine Learning Modeling

作　　者：罗鹏宇 胥小波 罗怡[1] 刘明春 LUO Peng-yu;XU Xiao-bo;LUO Yi;LIU Ming-chun(China Electronics Technology Cyber Security Co.,Ltd.,Chengdu,Sichuan 610041,China;No.30 Institute,China Electronic Technology Group Corporation,Chengdu,Sichuan 610041,China)

机构地区：[1]中国电子科技网络信息安全有限公司,四川成都610041 [2]中国电子科技集团公司第三十研究所,四川成都610041

出　　处：《通信技术》2020年第12期3087-3093,共7页Communications Technology

摘　　要：为了解决网络中的口令脆弱性评估、口令混淆检测等问题,研究基于机器学习融合告警的口令脆弱性评估方法,从大量的样本集中学习概率语言特征和文本上下文特征,通过随机森林算法和逻辑回归算法构建机器学习融合告警模型,实现了口令脆弱性的智能评估功能。将机器学习算法引入口令评估领域,大幅提高口令的评估检测能力,能够准确识别时序变化和恶意混淆的口令。In order to solve the problems of password vulnerability assessment and password confusion detection in the network,the password vulnerability assessment method based on machine learning fusion alarm is explored.This method learns probabilistic language features and text context features from a large number of samples,and constructs machine learning fusion alarm model by using random forest algorithm and logical regression algorithm to realize the intelligent evaluation function of password vulnerability.Machine learning algorithm is introduced into the password evaluation field,which can greatly improve the ability of password evaluation and detection,and accurately identify the password with time sequence change and malicious confusion.

关 键 词：弱口令 概率语言 机器学习 融合告警 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]