预览加载中，请您耐心等待几秒...

基于Stacking的Android恶意检测方法研究.docx

预览
1/2
2/2

在线预览结束，喜欢就下载吧，查找使用更方便

下载提示 文本预览

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

基于Stacking的Android恶意检测方法研究 基于Stacking的Android恶意检测方法研究 Abstract: Withtheincreasingpopularityofmobiledevices,Androidhasbecomethemostwidelyusedmobileoperatingsystem.However,therapidgrowthofAndroidapplicationshasalsoledtoanincreaseinthenumberofmaliciousapps.Asaresult,theneedforeffectiveAndroidmalwaredetectionmethodshasbecomeincreasinglyimportant.Inthispaper,weproposeastacking-basedAndroidmalwaredetectionmethod,whichcombinesmultiplebaseclassifierstoimprovethedetectionaccuracy. 1.Introduction Androidmalwarehasbecomeasignificantthreattosmartphoneusers,astheycanstealpersonalinformation,engageinfinancialfraud,andeventakecontrolofthedevice.Traditionalsignature-baseddetectionmethodsarenoteffectiveagainstnewandpreviouslyunseenmalware.Therefore,machinelearning-basedapproacheshavebeenwidelyadoptedtodetectAndroidmalware.Amongtheseapproaches,stackinghasshownpromisingresultsinimprovingtheclassificationperformance. 2.RelatedWork VariousmachinelearningalgorithmshavebeenproposedforAndroidmalwaredetection,includingdecisiontrees,randomforests,supportvectormachines,andneuralnetworks.Thesemethodsoftenhavetheirlimitations,suchasoverfittingorhighfalsepositiverates.Stackingisamodelensemblingtechniquethataimstocombinemultiplebaseclassifierstomakeafinalprediction,whichcanovercomethelimitationsofindividualclassifiers. 3.ProposedMethod Ourproposedmethodconsistsoftwomainsteps:baseclassifiertrainingandmeta-classifiertraining. 3.1BaseClassifierTraining Inthisstep,wetrainmultiplebaseclassifiersusingdifferentfeaturesandalgorithms.Weextractvariousfeaturessuchaspermissions,APIcalls,andopcodesequencesfromtheAndroidapps.Thesefeaturesarethenusedtotraindifferentclassifiers,includingdecisiontrees,randomforests,andsupportvectormachines. 3.2Meta-ClassifierTraining Aftertrainingthebaseclassifiers,weusethepredictionsoftheseclassifiersastheinputtothemeta-classifier.Themeta-classifieristrainedtocombinethepredictionsfromthebaseclassifiersandmakeafinaldecision.Weusestackingwithlogisticregressionasthemeta-classifierd