Ontology-based Heterogeneous Data Integration Research In Data Warehouse

The aggregate data is the basis of analysis and data mining when the data warehouse is building, and the data in warehouse must be sharing and reusing, can not be the new isolated data island. But the ETL scheme can aggregate data only on the level of systax and structure now, it can not solve the problem of data sharing and reusing and aggregating in semantic level, the ETL process is complex and have no intelligence, and it have a large workload. So this paper discusses an ontology-based approach of semantic aggregation in data warehouse, and solved the semantic heterogenous aggregation problem in stress.The framework is proposed: Firstly, the domain ontology is built, and the globe ontology of data sources are built with the guidance of domain ontology, after building the local ontologies of data sources, the mapping relations are got by the mapping arithmetic between the local ontologies and globe ontology. Through the ontology inference to get the connotative semantic relations. Finally, with the guidance of the final semantic mapping relations, the process of data extract, transform, and loading is started, and the problem of data intergration in semantic and intelligent degree in data warehouse is solved. On the aspect of domain ontology, a domain ontology of railway freight CRM is accomplished, it used OWL commented by W3C as express language to describe.Based on the research, OntoETL, as an Ontology-based ETL tool, is proposed, and according to the domain of the DW system in railway freight CRM, we designed and implement the data warehouse by the approach of ontology-based heterogenous integration.