欢迎来到天天文库
浏览记录
ID:36607162
大小:7.01 MB
页数:143页
时间:2019-05-12
《面向领域的语义信息检索与支撑技术》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、天津大学博士学位论文面向领域的语义信息检索与支撑技术姓名:郭炜申请学位级别:博士专业:计算机应用技术指导教师:何丕廉20061201ABSTRACTThispaperisbasedontheconceptsandtechnologiesofsemanticweb,whichisappliedtoinformationretrievalonspecialarea.Thispapermakesaresearchontheinformationretrievalbasingontheconceptso
2、fsemanticwebanditstechnology.Withinthepaper,relevantvocabulariesandrelationshipsinstockareawasabstractedanddefinedaccuratelybymeansofontologytechnologytobuildStockOntology.Also,stockinformationretrievalsystemhasbeendevelopedwiththefollowingthreecorem
3、odules:stockontologyfortheuserqueryextension,commonretrievaltechnologyforcoreretrievalmoduleinsystem,semanticrankalgorithmforresultrankingmodule.Meanwhile,self-adaptivehotpotisusedtoimprovethesemanticalgorithmandtheframeworkofsemanticwebinformationse
4、rviceisalsopresentedinthepaper.SemanticwebiscurrentlypopularservedasthemainstreamofnextgenerationofInteracttechnology,which,whenusedininformationretrievalbasedonspecialarea,canimprovetheaccuracyofretrievalresult.Ontheotherhand,ontologyisalsofittoresu
5、eknowledgeandimproveinteroperabilitybetweenmachinesautomatically.Sointhispaper,relatedtechnologiesofsemanticwebandtheconceptofdescriptionlogicwasintroducedandthenStockOntology(soinbrief)wasbuiltandcodedbyOWL.BasedonconstructionofSO,SemanticInformatio
6、nRetrievalSystembasedonSpecialArea(SIRSSAinbrief)ispresentedanddeveloped,whichcanprovidesemanticqueryextensionfunctiontohelpuserstoproducequeryeitherinstandardsystematicconceptualpatternorintriplenodeway.Aftertheresultsetbeingobtained,semanticranking
7、algorithm—OntoRankWasusedtoranktheresultdocumentset.Thealgorithmcananalyzetheconceptsorrelationshipsincertaindocumentandcalculatetheweightofeveryiteminresultsetwhichisveryhelpfultore—ranktheretrievalresultsemanticallyandprovideusersthemostrelevantret
8、rievalset.Toextendtheabovealgorithm,theconceptofself-adaptivehotpotwaspresentedinthispaper.Itcanbeusefultocollectinformationaboutuserclickfrequencyandreflecttheinformationorwebpagesusershavepaidmoreattentionto.Byuseofthisway,investorscouldconvenientl
此文档下载收益归作者所有