Design And Implementation Of Material Knowledge Extraction System Based On DBpedia

DBpedia is a comprehensive Linked Open Data of cross-domain and cross-language, and it has hundreds of millions of triples. DBpedia provides an abundant knowledge graph. However, DBpedia knowledge graph contains large amount of data of various discipline, and it is inconvenient for users to use metallic materials. The knowledge of DBpedia in the field of metal material is not comprehensive, and it lacks some data, for example, the performance data of metallic materials, the metallic materials specialist and literature, and so on. In order to structure metal material knowledge graph, this paper proposes a method based on DBpedia and design a system which can extract metallic materials domain data from DBpedia and extract the relevant metal domain data from Wikipedia web pages. In order to generate metallic materials knowledge graph, the research work of this paper is as follow:1) A stepwise extraction strategy is proposed to extract metallic materials data from DBpedia based on the structure of DBpedia. In order to extract metallic materials, we first screen through direct linked sub-graph semantic distance algorithm(DLSSD). And then we use the background knowledge algorithm to do further screening to extract all relevant metallic material knowledge.2) We process metal material performance table of Wikipedia website through form extraction algorithm based on ontology. In order to enrich metallic materials data, we link the performance data which is extracted from tables and the entities which are extracted from DBpedia together.3) The metal material references, materials experts, pictures and other information of Wikipedia web pages are extracted through keyword and template algorithm. Relevant data further enriches metallic materials data.4) Metallic materials knowledge graph and metallic materials ontology can be linked through the correlation framework of SILK and association rules, so that it can semantically enrich the metallic materials knowledge graph.The metallic materials extraction system based on DBpedia can extract metallic materials data from DBpedia and also extract related information from Wikipedia web pages to enrich metallic materials knowledge graph. The connection of metallic materials knowledge graph and metallic materials ontology can also be implemented in this system. This paper provides a method which can assist the experts to extract metallic materials knowledge from linked open data. The rich metallic materials knowledge graph is also provided in this paper.