Research On Service Oriented Semantic Retrieval Technologies Of Enterprise Data

With the appearance of the "Digital Enterprise", a number of data sources were accumulated in enterprise. How to find the relations in the enterprise information systems, how to achieve the semantically coherent data view, how to realize the semantic retrieval about enterprise data, how to offer convenient services via Internet/ Intranet, which are the problems the modern enterprise encounters. They are difficult to realize by traditional data integration technologies.Based on the discussion above, an new approach was proposed called Service-Oriented Semantic Retrieval on Enterprise Data. With the aid of the reasoner, based on sharing semantic model the query system including semantic information was built. Finally, a service-oriented architecture is carried out to make it loose-coupled and reused easily. The following achievements have been attained in this thesis.Developed the Semantic Retrieval Model After the investigations and research on the real world, this article proposed an approach which constructed domain semantic retrieval model based on the definition-tree first. Then on the basis of definition-tree, the semantic retrieval model which is the precondition of semantic retrieval operations was built. Semantic retrieval model is equal to the Formalization description to the enterprise system and problems and the semantic clashes were erased.Inference based on Description Logic The semantic retrieval model cares more about how to describe the roles on semantic within heterogeneous databases without consideration of supporting the inference. Hence, we make use of description logic which was good at inference on concepts, roles and instances to describe the model further, to realize the semantic retrieval to enterprise data on the basis of reasoner(Racer).Service-Oriented Semantic Retrieval This thesis carried out a service-oriented architecture to develop loose-coupled, easy extended and operated data integration system.This architecture services clients via cooperation of services of data semantic retrieval service and database wrapping service. On the course of operation, user needn't to understand how the system works.Prototype System Validation A prototype system of Service-Oriented Semantic Retrieval in Enterprise Heterogeneous Databases was developed on the basis of authorized or open-sourced software and all the above key technologies have been validated.