基于樹形結(jié)構(gòu)的Web信息抽取技術(shù)研究.pdf

1、廈門大學(xué)碩士學(xué)位論文基于樹形結(jié)構(gòu)的Web信息抽取技術(shù)研究姓名：任仲晟申請學(xué)位級別：碩士專業(yè)：計算機軟件與理論指導(dǎo)教師：薛永生20070401AbstractAbstraotWiththerapiddevelopmentofIntemet，Webisbecomingavast，distributed，andsharedinformationresourceMostofWebdataareintheformofHTMLDuetothesem

2、i—structurednatureofHTMLpages，WebpagesareeasyforexploringbyhumanbeingswhileitisdifficultforapplicationstoprocessandusethedataintheWebpagesTostrengthentheavailabilityofWebdata，providingmorevalueaddedservices，Webinformatio

3、nextractiontechnologycomesout，whichwrapstheWebresources，extractssemi—structureddata，andprovidessupportstoapplicationsusingWebdataTherefore，theresearchofWebinformationextractionisoneofthehottestresearchareasindatabasefiel

4、dandhasapromisingfutureInthispaperwefirstbrieflyintroducesomebasicconceptofWebinformationextractionandalsogiveashortintroductiontothedevelopmentofthetechnologyofWebinformationextractionThenwedescribethedefinitionoftheweb

5、pagesusedbyOuralgorithmSecondlywedescribe，compare，andanalyzeseveralkindsofWebinformationextractionmethodscommonlyusedatpresentindetail，pointingoutadvantagesanddisadvantagesofeachmethodFurthermore，wediscussthefuturedirect

6、ionofresearchanddevelopmentofWebinformationextractionFinallyweproposetreestructurebasedWebdataextractionalgorithminviewoftheinadequaciesoftheexistingmethodsOurtreestructurebasedalgorithmincludes：thealgorithmofHTMLtreecon

7、struction，thealgorithmofdataregionmining，thealgorithmofdatarecordmining，andthealgorithmofrecordschemagenerationOuralgorithmcleanstheWebpagesusingthepositioninformationofpageelements，minesdataregionbyhierarchicalclusterin

8、g，andgeneratesrecordschemafinishingdataitemextractionthroughtreematchingTheoreticalanalysisandexperimentalresultsshowthatouralgorithmcanimprovetheaccuracyandefficiencyofWebdataextractionKeyWords：Webdataextraction，Webmini