Research And Implementation On The Ontology-based Heterogeneous Data Exchange System

With the more and more popular of computer technology,and the rapid development of Internet technology,the desire for data sharing between enterprises or enterprise departments is becoming more and more intense.However,the most data sources are heterogeneous,and they can't be exchanged directly.The enterprises' information becomes some "information isolated islands".In order to solve the problem of information isolated islands and to ensure that the data and information can be shared and exchanged between enterprises,the key is to establish a unified, secure and stable platform for data exchange and a model of data exchange based on eliminating the heterogeneous data.As the ontology technology can determine the precise meaning of conception by using the strict definition of concepts and the relations between concepts,it has the potentiality of describing the source data semantics and solving semantic heterogeneous and can be used to express the knowledge which is universal approved and sharable.In this thesis,the ontology technology,the data exchange and integrate technology were reviewed,and an ontology-based data exchange method was introduced,which uses ontology as a tool of eliminating semantic heterogeneous.The main work of this thesis is illustrated as follows:1)This thesis chooses an appropriate method and an appropriate ontology description language to construct an overall ontology manually,based on the study of some existent methods how to constructe an ontology.So the overall ontology can provide a sharing lexical repository and the relationship of terminology in the lexical repository for exchanging data.2)This thesis studies three methods how to create an ontology from a relation database,as well as analyzes the advantages and disadvantages of the related tools, and implements the own tool what create an ontology from a relation database with the pre-defined rules method.3)The table and filed names may be described by Chinese and English mixed when a relation database was designed,so the local ontology which this thesis automatically created may be in Chinese and English mixed.To this question,the thesis implements mapping between Chinese ontologies and mapping between Chinese and English mixed ontologies with method based on the Linguistics and method based on Structure,after the comparative study of three methods how to map between ontologies.4)On the query interface,the SQL-like query languages,such as SPARQL,OQL, XQuery,SQL,are used,but it has certain requirement for its user.In this thesis database natural language interface is introduced into the data exchange system,and it can be used with few training.5)On the implementation of this data exchange system,the ontology is modelled by the OWL ontology language and manipulated by Java2 development kit.And then, the method proposed in this thesis was verified by a typical test case.