The Research Of Heterogeneous Data Integration And Fusion Based On Ontology

Integration of distributed heterogeneous data resources, implementation of data sharing, support for the user's management and decision-making with a more complete and reliable data service, have become the most important topic and direction of that deepen the construction and development of information systems in every field at present and in the future. Extracting data from various distributed and heterogeneous data sources, and processing data integration and data fusion is a fundamental task of the data integration.Based on the laborious research of existing data integration and data fusion method, combined with the project of our laboratory that is "Data Integration System Based on Data Service Matching ", this paper re-optimized design and rewritten the section on "heterogeneous data integration and data service management", and added a module processing data fusion for data from different data sources. Major work and contributions are presented as follows:1. After well studying domestic and foreign existing primary processing methods of data integration and data fusion, proposing a knowledge fusion method based the key attributes, applying this method into the data integration and management system, and improving the rationality of automatically merging the heterogeneous data.2. Re-optimized the design and rewritten the platform on "heterogeneous data integration and data processing services management" that the team have implemented : using Hibernate technology, rewritten the relevant code of the section about heterogeneous data access and query; using Spring technology, redesigned and rewritten the whole middle layer of this platform; optimized and improved front-end user interface design of the platform; added the module of automatically merging the data extracting from two different data source.Now the "heterogeneous data integration and data service management" platform can preliminary work, not only define the requirements of data integration with an intuitive and visual way, packet the flow of extracting data from multi-source, transforming data, consolidating data and integrating data into a data integration processing specification package (DIPSP) by operating the graphic object; but also process a data service cell package, registration, execution and response to the data service on the DIPSP.