Fuzzy Concept Lattice Gluing Based Fuzzy Ontology Merging Approach Research

Semantic Web is an extension of the World Wide Web, aiming at solving the problem of lack of semantics of current Web, by describing the information with a machine-understandable way. The Semantic Web uses ontology to describe information, express the meaning of concepts and the relationships among concepts. However, the information of real world is imprecise. This paper combined fuzziness and ontology to describe information of real world more accurate. In the ontology construction, due to different development teams and different understandings of knowledge, it’s easy to cause the problem of ontology heterogeneous. To solve this problem and achieve the goal of sharing and reusing resources, ontology should be integrated in some way. Ontology merging is the most effective way of ontology integration.However, ontology is still not good enough to reveal the intent and extent of concept. Formal concept analysis is a good way to define the meaning of a concept with its intent and extent and the relationships between concepts. But there are disadvantages in the aspect of coding and reasoning. So, combining ontology with Formal concept analysis can be a good way to express knowledge more accurate. A methodology of fuzzy ontology merging based on fuzzy concept lattice gluing is proposed by combining fuzzy ontology with fuzzy formal concept analysis.Fuzzy ontology is transferred into fuzzy concept lattice, and many operatings are conducted on the lattice such as similarity calculation and lattice gluing, and ultimately achieve the fuzzy ontology merging purpose. Sometimes, ontology may become very large, requiring modular it into a number of high cohesion and low coupling module, to be better sharing and reuse. A lattice partitioning method is proposed witch corresponds to generate modular ontology. Semantic similarity, structure similarity and lattice similarity of concept node from the lattice is calculated and integrated, improving the accuracy of merging process. According to the structural characteristics of ontology, a method of fuzzy ontology storage is proposed based on MongoDB database. MongoDB is a NoSQL database based on document storage, and it different from other database that widespread using today. It’s schema-free and document based, and fits to store information with object structure. So, it’s convenient to store ontology by transforming ontology to BSON structure. A prototype system named FOMS (Fuzzy Ontology Merging System) is designed and implemented. In order to verify the performance of proposed model and prototype system, experiment is conducted by using ontologies provided by OAEI2011. The results show that the proposed fuzzy ontology merging system has high accuracy and it has great value in self-learning of ontology since it can discover implicitly concepts automatically.