Research On The Storage Of OWL Ontologies Based On Graph Databases

The development of ontology technology promotes the expansion of ontology application domain.Intuitively,researches on ontology technology gradually moves from theoretical exploration to ontology application.However,in practical application,the size of ontologies is much larger and the complexity of ontologies is much higher than that in the laboratory scenario.In addition,users' query requirements are also variable.That leads to new challenges for the application of ontology storage and query.Researches on ontology storage become extremely meaningful and in urgent need of attention.The revolution of storage technology triggered by big data brings new opportunities to the storage of ontology.In this study,we combined a typical NoSQL-graph databases with ontology technology to explore whether ontologies can be stored in graph databases and if the answer is yes,how can they be stored.We use the Description Logic Theory as clues to build up a storage schema based on the graph database.Next,we designed the mapping rules in detail and proved the validity of the schema theoretically.Then we take the Global Product Classification as an example and build a product classification ontology using Neo4j in the guide of Relational Concept Analysis.At last,this study designed an ontology query experiment,which compared the query efficiency of the Neo4j ontology and OWL ontology we built.The result of the experiment proves the high efficiency of the ontology storage model proposed in this paper.This paper draws some valuable conclusions as follows.First,the development of NoSQL provides a new opportunity for ontology storage and among NoSQLs,the structure of the graph database is the most suitable for the storage of OWL ontologies.Second,the conceptual structure and instances in ontology have different semantic features,so they should be distinguished when designing the storage model.Third,Description Logic is the theoretical basis of OWL ontology,considering how to store the basic concept in Description Logic is the key to the design of ontology storage model.Forth,introducing the Relational Concept Analysis into ontology construction is an effective extension of ontology construction based on Formal Concept Analysis.Fifth,the ontology storage method based on graph database is superior to OWL document storage in query efficiency,which is an efficient and practical ontology storage method.The contribution of this study is that:(1)This study proposed a graph database storage schema for OWL ontologies and verifies its effectiveness.(2)This study fullfilled the ontology gr aphical database storage application using RCA,which contributes to the semi-automatic construction of ontologies.(3)In this study,a functional evaluation scheme of ontology storage method is proposed.Using this evaluation schema,we verified the efficiency of our storage scheme.