基于貝葉斯網(wǎng)絡(luò)的入侵檢測(cè)模型分析與研究.pdf

1、湖南大學(xué)碩士學(xué)位論文基于貝葉斯網(wǎng)絡(luò)的入侵檢測(cè)模型分析與研究姓名：李柏生申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：軟件工程指導(dǎo)教師：林亞平梅玖紅20070928莖三墨些翌墾釜竺金堡堡!!絲竺坌塹至至圣AbstractWiththedevelopmentofcomputertechnologyandthepopularityofinternet，thenetworksafetybecomesmoreandmoreimportantthemanagement’

2、Sprimarytargetofintrusiondetectionisthesafetywhichguaranteesallcomputer’Ssystems，networksystemandthewholeinformationstructureBaysiannetworkisnotonlyamethodoftheindeterminationdeductionbasedontheprobability，butalsoisthema

3、intoolwhichhandlestheindeterminationinformation，ThispaperappliesthecharacteristicsoftheBaysiannetwork，investigatesforintrusiondetectionsystemFirstly，thepaperanalyzestheNavieBaysinaNetwork(NB)indetailthemethodregardsallat

4、tributesinthetrainingsampleassubnodesoftherootnodethesub—nodesaremutuallyindependent，InthecurrentNBstructure，theNBacquiresthenetworkparameteronlythroughthetrainingsamplecomputingtheprobabilityofrootnodeandtheconditionpro

5、babilityofthesub—nodesintherootnode，inthesametimeitraisesefficientlythedetectionrateAlthoughtheNBhasthecharacteristicsofhigherdetectionrateitexiststhelimitsofsamefunctionamongallattributesTherefore，thepaperproposesanewre

6、visionforNB，whichisthemodelbasedontheChisquaretestThemodelconsidersdifferencesofattributesfuction，itdeletestheredundancyattributesanddoesnotconcernthem，SOCansimplytheNBstructure，theefficiencyisraisedindetectionsystemTheC

7、hisquaretestCanpromisestoacquiretheexactcharacteristic，andreducethecalculationquantity，inthemeantime，themodelraisestheclassifiedefficiency，andmakestheintrusiondetectionhaveabetterefulnessandadaptabilityIntheend，thispaper