Research And Application On Domain Science Data Semantic Mapping

Semantic Web and ontology play an important role in various research fields today, butthey are rarely used in the materials science domain. Materials science research oftenrequires a lot of data query. The high integrated linked-data of Semantic Web can exactlysatisfy this demand. To achieve semantic integration network in the field of materials,materials data must be converted from the relational database (RDB) to RDF as instanceof the domain ontology. In the practical application the materials data is so large that it isnecessary to recommend the mapping automatically.Nowadays highly integrated domain knowledge is a necessary tool for scientificresearch. In this thesis integrated knowledge network based on the domain knowledge isestablished in materials field as an example. First of all, the method of mapping frommetal materials to domain ontology is proposed. The proposed method refers to the latestW3C standard DM and R2RML. R2RML mapping documents (TTL files) are used torecord and organize mapping results after analyzing mapping of elements in the RDB andthe ontology. An algorithm which simulates the R2RML processor is designed to processthe TTL files so as to finally generate RDF. In the practical application the data inmaterials database is so large that mapping simply by manual method is not acceptable.This paper draws on the pattern of many famous knowledge base, creating SFC(top-level domain ontology framework) and MMMBB (Metal materials mappingbackground base) both domain knowledge base, and achieved the automatic mappingalgorithm based on the domain knowledge base, also proposed a new method of reverselyreasoning mapping by the database instance, greatly enhanced the mapping efficiency.Finally, this thesis presents a metal materials data mapping system based on R2RML anddomain knowledge base to achieve each of these functions and in this system theknowledge base can be edited and expanded. Meaningful research and exploration aboutmaterials informatics and the Semantic Web are presented in this thesis.