Acquiring ER Model By Ontology Reasoning Techniques

Database has been affecting people's daily life and facilitating the storage and sharing of data.In the era of the expansion of volume data,how to effectively manage different kinds of information is the core work in information system,which is also an important direction of our research.In the practical application,various problems will result in difficulties in designing database,such as lack of domain knowledge for database designers,ambiguous user requirements,differences in the communication between users and database designers and so on.The entity relationship model(ER model)regarded as a tool of conceptual models can abstract information in real life using graphical representation.ER model can intuitively express concepts and relationships between concepts and effectively solve those problems in the process of database design.Meanwhile,designing a comprehensive ER model is the core work in database design.With the endless development of network technology,domain experts upload domain knowledge to the network in order to facilitate the usage of it.People can achieve the interconnection of information through the opening and sharing network.Domain ontology can not only formally represent domain knowledge,but also be used for reasoning.We can obtain the desired information from domain ontology by ontology reasoning technologies and rules.In this paper we propose a novel approach for database construction based on the rich ontology resources and domain knowledge.With users' requirements we obtain ER model using the technology of logic reasoning combining with custom algorithms and rules.In this paper,we will introduce how to obtain ER model from ontology knowledge using ontology inference techniques,and validate the proposed method through a universal ontology univ-bench.owl.In addition,in order to verify the applicability of the method proposed in this paper,we have randomly chosen several different ontologies in different fields to carry out experiments.Using ontology inference techniques,we can not only obtain explicit domain knowledge but also contain hidden knowledge which provides rich data sources to construct ER model.On the one hand,with this method proposed in this paper,designers can make up the lake of domain knowledge.On the other hand,we have always been to users' requirements as the starting point,which will ensure that the ER model meets the actual requirements and reduce the tedious process of repeated communication between designers and users.The main work in this paper can be listed as follows.(1)According to the requirement terms given by users,we look ontology resources up on the network using the ontology query tool and regard it as the metabase.Reasoning machine is called to obtain the hierarchy relationships between concepts and to preprocess the metabase.(2)On the basis of the concepts set obtained by preprocessing,the entities used for constructing ER model will be acquired by the operation module of the recommended entity set.This module mainly includes three processes.Firstly,the concepts set will be extended to obtain the initial entities set init Entities using the algorithm combining with the ontology reasoning techniques.Second,a part of concepts in the initial entities set will be removed with our custom deletion rules to obtain candidate entities set candi Entities.Finally,the isolated concepts in the candidate entities set will be modified by the custom modification rules to obtain the final entities set of the ER model.(3)With entities set and ontology reasoning technology,relationships set used for constructing ER model can be obtained.In order to make the final ER model simply and clearly,we modify the relationships set in which exists inclusion relations between relationships according to the custom revising rules.(4)This paper obtains the relationship types and attributes with ontology reasoning techniques and introduces the situation that the sub concepts sub Class inherit attributes from super concepts super Class.With this way the simple ER model that only contains entities and relationships is enriched,on the basis of ensuring users requirements,which makes the semantics of the ER model more abundantly.