A Novel Dynamic Android Malware Detaetion System With Ensemble Learning

本文提出了一种名为EnDroid的动态分析框架,该框架利用多种类别的动态行为特征来实现高效的Android恶意软件检测。通过提取系统级行为踪迹及应用程序级别的恶意行为特征,结合特征选择算法去除无关特征并采用集成学习算法进行分类,实现了对恶意软件的有效识别。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

关键字:动态 Android恶意软件检测 集成学习
论文来源:2018 IEEE Access
简介:目前提出有效的Android恶意软件检测方法以阻止恶意软件的传播成为一个急切的问题。目前识别大规模恶意软件的主流倾向于通过动态或静态分析方法结合机器学习机制提取各种特征。一般静态分析方法在面对采用复杂混淆技术(加密动态代码加载)的恶意软件时变得无效。然而动态分析方法适合于解决此类恶意软件。本文提出动态分析框架EnDroid,基于多种类的动态行为特征实现高精度的恶意软件检测。涉及的主要工作如下:
  1. 提取特征,包括系统级别的行为踪迹、常见的应用程序级别恶意行为例如个人信息窃取、高级服务订阅和恶意服务通信。
  2. 特征选择:EnDroid采用特征选择算法去消除干扰的和无关的特征,提取关键行为特征。
  3. 构建EnDroid。通过对运行时监控提取特征,采用集成学习算法构建EnDroid,识别恶意软件。

发现:Stacking实现了最佳的分类性能,并且在Android恶意软件分类检测方面很有前景。
数据集:(1)良性:Google Play Store ;AndroZoo(2)恶意:Drebin (Drebin:Effective and explainable detection of Android malare in your pocket :NDSS 2014)
This book is based on our years-long research conducted to systematically analyze emerging Android malware. Some of our earlier research results and findings were reported in an IEEE conference paper entitled Dissecting Android Malware: Characterization and Evolution, which was presented at the IEEE Symposium on Security and Privacy (often mentioned as Oakland conference in the security community) in May, 2012 [77]. During and after the conference, we were pleased to receive and hear inquiries from colleagues with encouraging comments on the systematization of knowledge work that has been conducted in our conference paper. Partially because of that, we are motivated to expand our work and hope such efforts will be of service to the security and privacy community. Further, as part of that, we have released corresponding malware dataset for our study under the name Android Malware Genome Projectto the community. With that, we want to take this opportunity to thank our collaborators, Dongyan Xu, Peng Ning, Xinyuan Wang, Shihong Zou, and others, whose valuable insights and comments greatly enriched our work. The authors are also grateful to colleagues in the Cyber Defense Lab at NC State University, especially Tyler Bletsch, Zhi Wang, Michael Grace, Deepa Srinivasan, Minh Q. Tran, Chiachih Wu, Wu Zhou, and Kunal Patel. Special thanks also go to Susan Lagerstrom-Fife and our publisher for their great help and patience! This research was supported in part by the US National Science Foundation (NSF) under Grants 0855297, 0855036, 0910767, and 0952640. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, for the NSF. 1 Introduction ........................................ 1 2 A Survey of Android Malware........................... 3 2.1 Malware Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2 Malware Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2.1 Malware Installation . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2.2 Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.2.3 Malicious Payloads . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.2.4 Permission Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3 Case Studies ........................................ 21 3.1 Malware I: Plankton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.1.1 Phoning Home . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.1.2 Dynamic Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 3.2 Malware II: DroidKungFu . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.2.1 Root Exploits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.2.2 Command and Control (C&C) Servers . . . . . . . . . . . . . 24 3.2.3 Payloads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.2.4 Obfuscation, JNI, and Others . . . . . . . . . . . . . . . . . . . . 26 3.3 Malware III: AnserverBot. . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3.1 Anti-Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.3.2 Command and Control (C&C) Servers . . . . . . . . . . . . . 28 4 Discussion.......................................... 31 5 Additional Reading................................... 33 5.1 Books . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 5.1.1 Malware Detection and Defense . . . . . . . . . . . . . . . . . . 33 5.1.2 Smartphone (Apps) Security. . . . . . . . . . . . . . . . . . . . . 34 5.2 Conference and Workshop Proceedings . . . . . . . . . . . . . . . . . . 34 ix 6 Summary........................................... 37 References............................................ 39 Index ................................................ 43
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值