Research On Semantic Query Expansion Mechanism Based On Multiple Materials Ontologies

Along with continuous development of materials informatics,more and more data sources of metallic materials in different types are appeared,most of which contains a lot of materials knowledge.However,the storage structure and the naming of materials concepts are different in these data sources.The query results and the accuracy are often not satisfactory in these materials data sources by using the traditional keyword-based research methods.In order to improve the retrieval performance,this paper propose a semantic query expansion method based on multiple materials ontologies,which expands the seed query concept in these materials ontologies.So the query results are accuracy and contain a lot of relevant knowledge by this method.The main research work is shown as follows:1)Concepts expansion based on multiple materials ontologies: first,a concepts expansion strategy called S-SMOA based on multiple materials ontologies is proposed in this paper,which uses the SMOA algorithm based on string and the hierarchical structure of ontologies to obtain the similar concepts of the seed query concept in materials ontologies.Then,a relevant concept expansion strategy called P-Jaccard is designed,which utilizes the set-based Jaccard algorithm and the object properties of ontologies to expand the relevant concepts of the similar concepts.2)Pruning and complementing of expansion concepts: first,in order to improve the accuracy of concepts expansion,a method of concepts pruning based on Leacock-Chodorow algorithm is presented to filter the concepts in expansion concept set.Then,in order to ensure the recall of concepts expansion,a complementing strategy is used to supplement the expansion concept set.This strategy use Leacock-Chodorow algorithm based on the structure of ontologies to expand the seed query concept,which can get some new expansion concepts for the expansion concept set to improve the integrity of concepts expansion.3)Query processing of expansion concepts in data source: first,a matching strategy based on SMOA algorithm is used to obtain the matching concepts of the data source for these expansion concepts.Then,a semantic query scheme based on multiple materials ontologies is designed.The SPARQL query statements can be generated dynamically and automatically through this scheme,according to the features of the RDF triples data and the query pattern which are chosen by users,to retrieve the associated knowledge for thesematching concepts.The query results can be displayed in two intuitively and dynamically ways which are text and visualization.The semantic query expansion mechanism based on the multiple materials ontologies,not only uses the SMOA and Jaccard algorithm,but also utilizes the hierarchical structure and the object properties of ontologies,which can get more relevant concepts.Moreover,in order to improve the accuracy and completeness of the concepts expansion,Leacock-Chodorow algorithm is used to prune and supplement the expansion concepts in this paper.So this paper provides a semantic query expansion method and a system based on multiple materials ontologies for materials users and experts.