D_EEM_一种基于DOM树的DeepWeb实体抽取机制.pdfVIP

计算机研究与发展 ISSN N 111777TP J ournal of Comp ut er Research and Develop ment 47 (5) : 858865 , 20 10 DEEM :一种基于 DOM 树的Deep Web 实体抽取机制 1 2 1 1 1 寇 　月 　李 　冬 　申德荣 　于 　戈 　聂铁铮 1 (东北大学信息科学与工程学院　沈阳　110004) 2 (东软集团商用软件事业部 　沈阳　110 179) (kouyue @i se . neu . edu . cn) DEEM : A DOMTree Based Entity Extraction Mechanism f or Deep Web Kou Yue1 , Li Do ng2 , Shen Derong1 , Yu Ge1 , an d Nie Tiezheng1 1 ( Col leg e of I nf orm ation S cience an d Eng i neeri ng , N ort heas tern Uni vers i ty , S heny ang 110004) 2 ( ) B us i ness S of tw a re D i v is ion , N eus of t Group L t d . , S heny ang 110 179 Abstract 　Wit h t he increa se of Web dat aba se s , acce ssin g Deep Web i s beco ming t he main met ho d to acquire infor mation . Becau se of t he lar gescale un st r uct ured cont ent , het ero geneo u s re sult and dynamic dat a in Deep Web , t here are so me new challen ge s for entit y ext raction . Thu s it i s import ant to solve t he p roblem of ext racting t he entitie s f ro m Deep Web re sult p age s eff ectively . By analyzing t he charact eri stic s of re sult p age s , a DOMt ree ba sed entit y ext raction mechani sm for Deep Web (called DEEM ) i s p re sent ed to solve t he p roblem of entit y ext raction for Deep Web . DEEM i s mo deled a s t hree level s : exp re ssion level , ext raction level , collection level . Therein t he co mponent s of region location and semantic annot ation are t he core p art s to be re searched in t hi s p ap er . A DOM t ree ba sed auto matic entit y ext raction st rat egy i s p erfor med in