資源描述：
《網(wǎng)絡(luò)流量異常檢測的小波分析方法》由會員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、電子科技大學(xué)碩士學(xué)位論文網(wǎng)絡(luò)流量異常檢測的小波分析方法姓名：高軍申請學(xué)位級別：碩士專業(yè)：通信與信息系統(tǒng)指導(dǎo)教師：胡光岷20060601摘要關(guān)鍵詞：異常檢測，多尺度，小波包，滑動時(shí)窗，無窮可分層疊ABSTRACTABSTRACTNetworktratncanomalyreferstothestatusthattrafficbehaviorsdepartfromthenormalbehaviors．Manyreasons，suchasthemisuseofnetworkequipments，networkoperationsanoma
2、ly，flashcrowd，networkintrusionandsoonwillCaUSenetworkanomaly．ThecharacteristicofanomalytrafficsisthatitbreaksoutwithoutanyomenandCandestroynetworksandcomputersinashorttime(Forinstance，theoutburstoftrafficbehaviorcausedbyspecificattackprogramsorwormburst)．Therefore，tod
3、etectanomalyrapidlyandaccuratelyisoneofthepreconditionofensuringtheefficientnetworkoperationanddetectionofanomaloustraffichasbecometheattractiveandvaluablesubjectinthepresentacademicandindustrialcircles．Thispaperintendstoresearchthewaveletpacketanalysismethodwhichcand
4、etecttheanomalyaccuratelyandreliably,andachievethereal—timeperformance；aimattheproblemoftrafficanomalydetection，researchthefastwavelet(paeket)transformalgorithmbasedonslidingwindow；aSthetrafficanomalymakeadifferencetothenetworktrafficIDCcharacteristic，researchandimple
5、mentthedetectionbasedontrafficIDCcharacteristic．(1)Atfirst，thispapersummarizesthenetworktrafficanomalydetectionmethods．Analyzingtheresearchactualitybothhereandabroad，wecategorizethenetworktrafficanomalydetectionmethods，pointoutthedeficiencyofthepresentresearches，aimat
6、theproblems，proposedsomenewresearchclewsandpossibletechnologylines．(2)r111ispaperpresentssomebasictheoryaboutwaveletanalysis，andproposesafastwaveletQacket)algorithmbasedonslidingwindow．Bystoringthepartofwaveletpacketcoefficientsinmemoryinadvance，Weavoidrepeatingcalcul
7、ationandgainafastwaveletalgorithmatthecostofincreasingsomestoragespace．nleexperimentsprovedthatthefastalgorithmcansufficientlysupportOurnetworktrafficanomalydetectionmechanism．(3)Aimingatthedeficiencyofpresentmethodsofnetworktrafficanomaly，thispaperproposesanewscale—a
8、daptivedetectionmechanism．Bymeansofwaveletpacketdecomposition，Ourmethodhasthesanledetectiveabilitytomiddleandhighfrequencyas
