Research On Lightweight Data Integration Method Based On Semantic

With the development and wide application of database and network technology, the datas of datasource show distributed and heterogeneous characteristics. Due to business needs, enterprises establish the appropriate computer processing system and storage a large amount of datasources. These datasources are isolated from each other and their structures are various, which leads to the generation of "Information Island". In order to make better use of these datasources serve enterprises in the decision-making and business processes, it is needed to integrate these distributed, heterogeneous datasource and implement data sharing. When the number of datasource increases, the cost of the existing method will become very high. So, how to use a smaller price to solve the syntactic, semantical and structural heterogeneity of the distributed and heterogeneous datasources has become a research hotspot.This article researched the lightweight data integration method based on semantics for the existing method which has the large integration quantity and high integration cost. Firstly, the article described the related technologies of data integration, and gave the lightweight data integration framework based on semantics. Then proposed a unified representation of metadata and further gave recognition algorithm of local metadata, ontology extraction rules and global metadata integration algorithm. Finally, verified the feasibility of the method through the experiments.This article completed the following work:(1) This article proposed a metadata formats to the need of the lightweight data integration.This metadata compared with the existing metadata has the following advantages: global uniqueness by URL naming, semantics by using RDF to express of the metadata and more comprenhensive datasoure information. Identified local ontology in the local datasource, and extracted local metadata using metadata extraction algorithm for each local datasource.(2) This article established the mapping between local ontology with the global ontology and proposed the global metadata integration algorithm based on the mapping between global ontology with local ontology and the mapping between local metadata with local ontologyang to get the global metedata which solve the syntactic, semantical and structural heterogeneity. Further, gave the update strategy of global semantical metedata to ensure real-time data integration.(3) Under the guidance of the above theories, designed and implemented a lightweight data integration system based on semantic using related technologies, which mainly consists of four modules: ontology recognition module, metadata extraction module, the ontology mapping module and the global metadata integration module.Finally, the artcle proved the feasibility and validity of this method by querying experiments.