Ontology Matching Based On Multi-strategy Concept Similarity Calculation

With the advent of the era of big data,data on the internet has exploded.In order to better use and share these knowledge,ontology is introduced into the Semantic Web because of its excellent performance in storing and describing knowledge.Due to the subjectivity of ontology in the process of development and the distributed nature of Internet,even in the construction of ontology in the same field,the ontology built by different organizations and individuals is different,so the ontology in the same field is different,which is called ontology matching.They are usually not interoperable,nor can they be understood between different users.In order to make full use of these ontologies to accomplish the task of knowledge integration and communication on the Internet,it is necessary to find out the matching relationship between these ontologies,that is,to eliminate ontology heterogeneity by using ontology matching technology.Ontology matching is an important technique to solve the interactive operation between different ontologies and to find the matching relationship between different entities.It is very important to realize the interaction between applications with different ontologies.Due to the explosion of the number of ontology in big data,it is urgent for some methods to perform the task of ontology matching automatically or semi-automatically,but the matching effect of ontology matching system needs to be further improved.A single ontology matching method considers only the elements specific to the ontology,and the matching effect is not satisfactory.This paper adopts ontology matching from multiple angles and implements ontology matching method.This article mainly contains the following contents:?1?Conceptual similarity calculation for the name of the ontology.a multi-strategy based ontology matching algorithm is proposed.The algorithm first calculates name-based similarity for two given ontologies,O₁ and O₂,based on string formula and WordNet dictionary method,and introduces annotation information..?2?Conceptual similarity calculation based on attributes,instances and structures.We use the sum product algorithm to calculate the case-based concept similarity.By filtering attributes,calculate the attribute-based concept similarity,and use the structure-based noisy-or model to calculate the similarity,and combine the values of the three algorithms to get the initial matching.The final match is obtained by repeated iterations.?3?Based on the algorithm studied,makes experiments on the method presented in this paper.By using the international standard evaluation data set and evaluation method provided by OAEI,the experiment shows that the algorithm has good performance,can effectively improve the accuracy and the recall rate,and has achieved remarkable results in the matching efficiency.