Metadata matching based on Bayesian network in DataSpace

The concept of Dataspace was proposed as a new data management paradigm to help unify the diverse efforts on flexible schema, data cleaning, and data integration being pursued by the database community. The heart of Dataspace is “pay-as-you-go” philosophy, which provides Best-Efforts service and brings more convenience to data management. To elevate the service quality, extracting the matching between metadata becomes a crucial problem, while the uncertainty of Dataspace hinders the metadata matching. This paper introduces the Bayesian network model to extract metadata matching, solves the problem of uncertainty by probability reasoning, and can complete the missing metadata as well. The method well suits the circumstance of restrict model nodes and well structured consequence of node.