Research On Linked Data Model Based On The Semantic Similarity Analysis

With the rapid development of Internet technology,the data resources in the network are increasing in full speed,which puts forward new requirements for the efficient sharing of resources in the network.In 2001,Tim Berners-Lee,the father of the World Wide Web,creatively proposed the conception of semantic web,which points out a new path for the interconnection and interoperability of data resources in the Web.Semantic similarity calculation,which is the foundation of the semantic web,has been widely applied in fields such as information retrieval,machine translation and personalized recommendation in recent years.Linked Data is an important stage of semantic web development,because it contains rich structured and interrelated data,which makes it possible to realize the semantic associations.In this thesis the related algorithms of data semantic similarity calculation are analyzed,and some improvements for the semantic similarity between different types of data are proposed.First of all,for the text type of data,the original algorithm just considered the effects of the node distance on similarity,we further consider the influence of the depth of nodes.We use the depth of the nearest common ancestor node of the two sememes nodes to represent the depth of the two nodes that need to compute the similarity.Secondly,in view of the RDF data,we consider the influences of the structure characteristic of the RDF triples and the semantic information of the nodes in the RDF graph on similarity.The influence can be represented respectively by the connected-edge weights and the nearest common ancestor node in the domain ontology model.The simulation results show that the above two new algorithms can effectively improve the accuracy of the results of the semantic similarity calculation.On the basis of the improved semantic similarity algorithms,the general procedure of the linked data modeling is detailed described.Combining with the characteristics of the specific books bibliography data,we add the linked associations in logic level and semantic level to RDF data describing bibliographic information,which make the isolated RDF metadata can be connected to each other,and then build up the linked data model based on the bibliography information.