摘要
随着Android恶意应用的数量越来越多,危害越来越大,提出一种行之有效的恶意应用检测方法也越来越紧迫。Android权限是静态检测中最为有效的特征,但由于其种类繁多,且每个权限对分类的贡献大小不一,因此,选择对分类贡献明显的权限尤为重要。针对此问题,论文提出了一种结合Relief算法和Apriori算法的特征选择方法--RApriori。该方法利用Relief算法对Android权限去冗余,进行一次特征选择;然后利用Apriori算法对去冗余后的权限关联挖掘,进行二次特征选择。最后,利用随机森林算法对经特征选择后的权限分类建模。通过实验验证该方法的有效性和可行性,实验结果表明,RApriori方法的恶意应用检测率达到90%,而误报率仅为9%。
With the increasing number of Android malicious applications,it is becoming more and more urgent to propose an effective detection method for malicious application.Android permission is characteristic of static detection of the most effective,but because of its variety,and each authority contributes to the classification of different sizes,therefore,choosing the contribution to classification of apparent authority is particularly important.Aiming at this problem,this paper proposes a feature selection method combining Relief algorithm and Apriori algorithm,that is RApriori.This method makes use of the Relief algorithm to make a feature selection of Android permission.Then,the Apriori algorithm is used to excavate the permission associated with redundancy and make a secondary feature selection.Finally,the model of permission classification by random forest algorithm is used.The experimental results show that the detection rate of malicious application of RApriori method is 90%,and the false alarm rate is only 9%.
作者
李秀
陆南
LI Xiu;LU Nan(School of Electronic Information,Jiangsu University of Science and Technology,Zhenjiang 212000)
出处
《计算机与数字工程》
2019年第12期3089-3094,共6页
Computer & Digital Engineering
基金
国家自然科学基金项目(编号：61601206)资助
