基于智能Agent的個性化Web瀏覽器研究與實現.pdf

1、江西師范大學碩士學位論文基于智能Agent的個性化Web瀏覽器研究與實現姓名：鐘茂生申請學位級別：碩士專業(yè)：計算機軟件與理論指導教師：王明文丁暉2003.5.1堊墮墮蔓盔蘭堡主堂垡笙壅墨王塑墼壘壁!!墮塵絲些!生型苧堡墮壅皇壅翌————壘墮竺竺AbstractTheIntemetespeciallyWWWhasmademuchprogressinthelastyearsTheAmericarlInteractcommitteeissue

2、daresearchreportaboutthestateofInteractofthe2∞Owhichsaid出atthenumberofW嵌)PageintheInteracthasarrivedat20hundredmilliontilltheSeptemberof2000andwillarriveat40hundredmilliontilltheearlyof2001Duringtheprocessofsearellinform

3、ationinthetremendousamountofvastanddisorderinformationwodd，manypeoplehavegradtha／lywilderedthedirectionHowtosearchquicklyandeffectivelytheinformationthatreflectsuserlsrequirementorhisinterestsandhowtotake蝴ladvantageofthe

4、sevaluableresoulcealetheaimofresearcher’sputsuingatalltimesManyresearchershavegenerallyacknowledgedthatpersonalizedinformationserviceisaverygoodapproachTorealizetheindividuationserviceofW曲informationweneedmanydomanialthe

5、oreticsandtechniques，forexamplemostlyInformationSearches，InformationFiltrationComputerNetwork，DistributedDisposalDataBaseManageSystem，DataMining，NumericLibrat3，DisposalofnaruralLanguage4rt湃cialIntelligenceMachineLearning

6、andsoonThustealizetheindividuationserviceofW曲informationwewillbeencounteredgreatchallengesandwehavetodoacompositiveandhardtaskWehaveundertakedsomestudytealizetheindividuationserviceofWebinformationinthispaperThegoalofoar

7、studyistostrengthenorimproveonthepowerofWebBrowseroftheclientcomputertoenchancethequalityofinformationgatherandthee掰cieneyofinformationsearchtodecreasetherepeatedsearchortransferofWebinformationtealizetheinformationsearc

8、hofuser，sinterest—Oriented，toincamateandreflecttheuser，srequestmentoftheindividuationinformationservice，byimplantingtheIntelligentAgentwiththeabilitiesofselfadaptedlearning，informationfiltrateandautomaticsearchTorealizet

9、heusersinterestOrientedindividuationWebinfcrmationservicewewiljmeetthefollowingquestions，whicharehowtodenotetheLiser’sjnteresthowtolealTltheuser’sinterestfortheagenthowtosearchtheWEBInformationofmeuser’simerestorientedan

10、dhowtofiltratetheInformationforusersinterestandsoonWeputforwardsomemethodstoresolvetheabovequestionsThefollowingareourresearchwerkFirstlywehaveputforwardamathematicfc’ITll一interestvector——toexpresstheuser’smterestandhave

11、givenaformulatocalculatequantificationallytheInterestdegreeofinterestitern，Secondlyweapplythemethodofretntorcementlearningtointelligentlearningagenttomakejt】earntheuserlsinterestmoreprecision，morequicklyandmoree歷ciencytr

12、lakeitdiscoverthedivertoftheuser’simerestintimeThirdlywehavedevelopedameanofinformationsearchandfiltrateonuser“sinterest。Lastb二wehavebroughtforwardaobjectbasedprogrammingschemeforAgent，namely，amethodtodesignAgentsystemby

13、extendingtheobjectoriebtedmodelingtechniqueormethodologywehavestudiedtheaboreproblemsanddesignedaPersonalizedWebBrowserBasedonIntelligentAgentandrealizedsomeimportantfunctionsThesystemhasthef01]owin￡tunctionsFirstlythesy

14、stemallowstheusertologininthesystemortegisterintothesystemanditwillcreatetheUser’3profilewhentheuserdidn’tregisterintothesysteminthepust，Secondjy^weoanbrowsethewebpagebyusingthesystemThirdly，theagentclnllearnsomeuser’sin

15、terestbyInteractionwithuserLastlythesystemCallfiltratesomewebpageinthelightofuser’sinterestprohieIhespecialtiesoft吣systemBlepcrsonalization，selfdeterminationmaintenancewebmtormationandknowiedge’base，collaboration，expansi

16、bilityandsoonThesystemisnotverypenectatthepresenttimeThus，wehavetesearchanddevelopthesystemulteriorlyint11e觸ureKeywords：WWW，Informati。nSearchInformati。nFiltrate，Pers。nalizati。n，IntelligentA齜MachineLearning，Q—LearningAlgo