Design And Implement A Knowledge Base Management System Oriented Linked Open Data

Semantic Web proposed by W3C aims to make machine understand words,concepts and their semantic relationships.The semantic datasets are always published as Linked Open Data(LOD).In this environment,people store knowledge and develop the intelligent applications by Knowledge Base.However,the most existing methods of knowledge management have no enough ability to support perfect Semantic Web's standards and deal with semantic relations,like a traditional database which only CRUD with general data.This paper designs and implements a Knowledge Base Management System(KBMS)oriented Linked Open Data(LOD)to help knowledge engineering or other semantic applications.The KBMS is a system based structured/unstructured LOD which support SPARQL and semantic query.It stores and manages data with dividing knowledge into two distinct parts.The instance data(RDF)will be stored in Virtuoso as Data Layer of Knowledge Base and the metadata(Ontologies)will be stored in Neo4J as Pattern layer.The article analyzes the standards of Semantic Web and the key points of Ontology Engineering and do a survey about knowledge base tools.Then five requirements of knowledge management are summarized.Based on these requirements,the KBMS is advised to consist of five layer named Parser,Model,Store,Query and Application.According to this design scheme,many function modules must be accomplished,including parsers&writers,datatypes,database,query engineer and ontology algorithms.Further,this paper illustrates the realization of the system's critical technology.At first,the system's datatypes must be designed to map to graph structure in Neo4J by Jena.It provides a foundation for implementation of parsers&writers.Both RDF data model and Ontology model can be represented by the datatypes.Secondly,the paper propose several key points about automation ontology management.Based on concept and concept hierarchy,the article proposes a comprehensive similarity measure for ontology search,and uses supervised learning classifier and WordNet dictionary system to implement ontology mapping and merging algorithm.Moreover,the analysis of semantic query is completed based on dependency analysis,and the query language conversion process of Multi-type Query Engine is designed.Finally,the paper proposes a number of test methods,using BSBM and other test datasets to verify the availability of the functional modules in all aspects.The achievement of this paper is to put forward a new scheme of knowledge base management.With clarifying the demand of knowledge base management,the paper designs and implements an available knowledge base management system,which provides a new idea for the development and innovation of knowledge engineering.One of inno-vations of this paper is to realize the separate management of Knowledge Base.In order to reduce the manpower cost of ontology management,the system use automatic algorithm,i.e.ontology search,ontology mapping and ontology merging.And it also provides many kinds of knowledge query ways.