语义检索中的概念相似度计算和关键技术分析-concept similarity calculation and key technology analysis in semantic retrieval.docxVIP

Abstractontologyintotwomainclasses:subsumptionrelationsandnon-subsumptionrelations.Weconnectsubsumptionrelationswithsimilaritycomputationandnon-subsumptionrelationswithrelevancecomputation.2.Westudyandexpandthecomputationmethodofstrengthfactor.Thenweoffercorrespondingformulasofstrengthfactorfordifferentrelations.3.Tosolvetheproblemofhowtosynthesizedifferentinfluencesofspecificrelationsonthesimilarity,wecomeupwithnon-correlationfactorforproperties.4.Thispaperputsforwardreversefactortodealwiththeasymmetricalprobleminsimilarityandrelevancecomputation.Attheend,weproposecompletealgorithmsincludingalltheimprovementsabove.Inchapter5,weaimattwocoretechniques--semanticpretreatmentandsemanticqueryexpansionofsemanticretrievalsystem.Weofferasemanticqueryexpansionalgorithm,andapplythesemanticsimilaritycomputationmodelherewhichisproposedinchapter4.Attheendofthechapter5,webuildasemanticqueryexpansionmoduleandrealizeallthealgorithmsweofferinthispaper.Thenwevalidatetheimprovementsthroughexperiments.Alltheworkinthispaper,establishesboththetheoreticalandthepracticalbasisforimplementationofsemanticsearchsystem.Keywords:ontology,semanticdistance,semanticsimilarity,semanticrelevance,semanticsearch,semanticqueryexpansion语义检索中的概念相似度计算和关键技术研究目录摘要···············································································································································IAbstract·········································································································································II目录··········································································································································IV第一章绪论11.1问题的提出及研究意义···································································································11.2研究现状··························································································································21.3本文的工作·····································································