Data quality of data warehouse is an important concept. Goal-Question-Metric is a model that adopt by former documents. And it is based on question asking to develop data quality of data warehouse.First, the paper reviews the former data warehouse's architecture and theory of data quality. Next, the paper proposes a three layers architecture which incorporate data quality factors. And a way of improving data quality are raised which satisfy the different layers of data warehouse to support DSS efficiently. The highlight in this paper describes a mechanism which is used to develop data quality in the architecture. And the mechanism include cluster and transfer agent. Base on the data source and requirement of data quality, a hierarchy is formed. During the forming of clusters, transfer agents are necessary which overcome the disadvantages of GQM. A knowledge search tool in transfer agent not only help to forming cluster but also classify data. A measure agent, optimizer and evaluator is included in transfer agent. End, we develop theory and ways evolution of date. A three layers architecture of data warehouse based on data quality is built.
|