OPTIMIZATION OF DATA WAREHOUSE MODEL ON THE BASIS OF ATTRIBUTE CLUSTERING

Resumen

Performance of information system is one of the main indices of its efficiency. Most existing information systems and software complexes operate with relational data warehouses. Design of data warehouses for information system is comprised of sequential development of conceptual, logical, and physical models. This design sequence has proved its efficiency in the case when the requirements to database, defined by domain area, are strictly formalized, the load on data processing center is forecastable. However, there are cases when several information systems operate with data warehouse. Each of them has its own requirements to information, sometimes contradictory, which should be presented by database in the shortest possible time. In such cases, forecasting of load, detection of the most resource consuming requests require for system analysis.