查询结果:   孙磊,韩静丹.基于BHNB的细粒度的Android恶意应用检测模型[J].计算机应用与软件,2017,34(10):310 - 315.
中文标题
基于BHNB的细粒度的Android恶意应用检测模型
发表栏目
安全技术
摘要点击数
659
英文标题
A FINE-GRAINED ANDROID MALWARE DETECTION METHOD BASED ON BHNB
作 者
孙磊 韩静丹 Sun Lei Han Jingdan
作者单位
信息工程大学三院 河南 郑州 450000     
英文单位
College of Cryptography Engineering, Information Engineering University, Zhengzhou 450000,Henan,China     
关键词
Android 动态分析 层次朴素贝叶斯 集成学习 恶意应用检测
Keywords
Android Dynamic analysis Hierarchical Nave Bayesian Ensemble learning Malware detection
基金项目
国家重点研发计划项目“协同精密定位技术”(2016YFB0501900);国防预研基金项目(910A26010106JB5201)
作者资料
孙磊,研究员,主研领域:云计算基础设施可信增强、可信虚拟化技术。韩静丹,硕士生。 。
文章摘要
为进一步提高Android恶意应用的检测效率,提出一种基于BHNB(Bagging Hierarchical Nave Bayesian)的细粒度Android恶意应用检测模型。该模型首先对样本库中的应用进行类别划分,并分别对其进行动态分析,提取各个应用程序的行为信息作为特征;然后,采用层次朴素贝叶斯HNB(Hierarchical Nave Bayesian)分类算法对各类应用特征集合进行分别训练,从而构建出多个层次朴素贝叶斯分类器;最后,采用Bagging集成学习方法对构建出的多个层次朴素贝叶斯分类器进行集成学习,构建出基于层次朴素贝叶斯的Bagging集成学习分类器BHNB。实验结果表明,该模型能够有效检测出Android恶意应用,且检测效率较高。
Abstract
In order to further improve the detection efficiency of Android malicious applications, this paper proposes a fine-grained Android malware detection model based on BHNB (Bagging Hierarchical Nave Bayesian). First, the model classified the applications in the sample database and dynamically analyzed them respectively, and extracted the behavior information of each application as features. Then, HNB (Hierarchical Nave Bayesian) classification algorithm was used to train all kinds of application feature sets respectively, so as to construct several layers of Nave Bayesian classifier. Finally, the multi-level Nave Bayesian classifier was constructed by using bagging ensemble learning method, building up the Bagging ensemble learning classifier based on Hierarchy Nave Bayesian algorithm-BHNB. The experimental results demonstrate that the proposed model can effectively improve the detection efficiency while improving the detection accuracy.
下载PDF全文