基于隱馬爾可夫模型的網(wǎng)絡入侵檢測系統(tǒng)研究.pdf

1、廣東工業(yè)大學碩士學位論文基于隱馬爾可夫模型的網(wǎng)絡入侵檢測系統(tǒng)研究姓名：趙玉明申請學位級別：碩士專業(yè)：計算機軟件與理論指導教師：滕少華20050426AbstractToday，Intemethasalreadybeena工1indjspensablepanofpe叩le‘sdaily1jfeasthecomputcrnetworkhasdeVel叩edr印idly；atthesametjme，thesec嘶tyofnetworkdraw

2、smoreandmo陀atte洲onofpeopJe。Atpresent，therearemanytechn0109iesofnem，orksafeguard，forex踟ple，fire、vall，accesscontrol，da詛encryption；Butallofthesetechnologiesarestaticlydefensivetools，whichcannottotallyassllI瞎thcsecurityofnet

3、workandresist血eattackofhackerUndertllisbackgmund，IntrusionDetectionSystemhasbeenanewdirection，whichcanini!tiatively跖ddyn鋤icallyprovidcsecuritysafeguarda11dsupplymelimitationtraditionalnct、vorksecuritytechnologiesTheHidde

4、nMarkovModelhaSmallyexcellentfeatufe，forinstanceitsmatufeaIgorithrn，hi曲e銜ciency，easytrai血ngetc；soithaseXtensive印plication抽a10tof矗elds，su曲asphoncticfecognitionThemddenMarkovModelcanreducethe蹦se—positiVerateandincreasemede

5、tectionrateinanomalyintmsiondetectionAtpresent，thedataresourceofa】101nalyintrusiondetcc灶onbasedonHiddenMarkovModelstemfbmthehostcomputer(forex鋤ple，systemcall)Thesystemhasobtained廿1egoode、叩erimentalresultNetworkintmsionde

6、tectionisusedtodetectintrusionbyc砒chingnetworkpacketanddrawingcharacteristics(IPaddress，poft，etc)Themaindifrcultyinme印pIicationofHidenMarkovModelinnetwork卸omalyintmsiondetectionistodecidetheobseⅣationsymbolsofHMMInthispa

7、per，wehaVeproposedthemethodtoconfirmobservationsymb01sofHiddellM甜koVModel，鉍d也enapplyHMMtonetworkanomalyintrusiondetectionOntllebasisofconfimlingmodelobservationsymbolsofHiddenMarkovModel，mispaperhasproposedanetworkanomal

8、yimrusiondetectionsystemmodelbasedonHMM，whichadoptsbotllmisusedetectionanda110malydetectionandisrealizedThelaStexpcrimemresultslowsmat也esystemllasrelativelyhighdetectionrateandcandetectunknownaHackundersomecircumstanceof