摘要
随着Android操作系统在智能设备上的广泛应用,Android应用的安全性检测成为了当前关注的重点。为了从Android应用程序中检测出恶意软件,研究Android应用静态分析技术、动态分析技术及基于机器学习的Android应用检测技术。提出一个通用的恶意软件检测框架。该框架通过逆向工程从Android应用中提取(安全应用、受感染应用)特征信息并建立关键信息特征库。通过机器学习建立检测模型,采用分类检测技术完成检测。通过该检测框架,可在软件安装前执行应用安全评估,其检测正确率高,并具有良好的扩展性,为Android应用的安全性检测提供参考。
With the wide application of the Android operating system in smart devices,the security detection of Android applications has become the focus of current attention.In order to detect malware from Android applications,this paper studied Android application static analysis,dynamic analysis and Android application detection technology based on machine learning,and proposed a general framework for malware detection.The framework extracted feature information from Android applications(secure applications,infected applications)by reverse engineering and established key information feature database.We established the detection model through machine learning,and completed the detection through classification detection.The framework can perform application security assessment before software installation.It has high detection accuracy and good expansibility.It provides a reference for the security detection of Android application.
作者
刘玮
李蜀瑜
Liu Wei;Li Shuyu(College of Mathematics and Computer Science, Chongqing Normal University Foreign Trade and Business College, Chongqing 410520, China;College of Computer Science and Technology, Shaanxi Normal University, Xi'an 710062, Shaanxi, China)
出处
《计算机应用与软件》
北大核心
2019年第6期322-326,共5页
Computer Applications and Software
基金
重庆师范大学涉外商贸学院校级教改项目(JG2016015)
关键词
安卓
安全性检测
静态分析
动态分析
机器学习
Android
Security detection
Static analysis
Dynamic analysis
Machine learning
