面向语义网的本体学习方法研究-计算机应用技术专业论文.docxVIP

中国科学技术大学硕士论文 中国科学技术大学硕士论文 面向语义网的本体学习方法研究 Abstract With the rapid development of the World Wide Web，the problem of the management and retrieval of the huge amount information emerged．The Semantic Web is such technology which is designed to overcome the bottleneck of the development ofthe WWW． By annotating the current information source with the semantic information，the Semantic Web extends current Word Wide Web，and makes the computers work with people automatically．As the semantic metadata of the knowledge in the Semantic Web，the ontology is the key element ofthe Semantic Web．The ontology is the formal explicit specification of domain conceptualization．It provides the domain．the common understanding of the shared knowledge． The construction ofthe ontology is very complex process，multi-domain experts needed in the process．Although some ontology engineering tools，such as Prot696， have been developed，the process of the ontology building is still time consuming and expensive task．This situation affects the rapid development o f the Semantic Web． Ontology leaming is the set of methods and techniques used for building ontology in a semi-automatic fashion using several sources．It be used by ontology engineers to construct ontology more easily．In this paper,we analyze the ontology learning method from different key tasks，the concept extraction，instance extraction and relation extraction． The main drawback of the current ontology learning methods is the lack of the semantic analysis of the information sources．To solve this problem we imported the frame semantics and FrameNet corpus into ontology learning method．Through shallow semantic analysis，we construct the syntactic and semantic mapping structure． Based on this structure，our method incorporate the semantic context of the corpus into the concept instance extraction and non．taxonomic relation extraction phase of the ontology learning circle．We also develop new ontology learning system named FraSEOL which combines the tra