Enhancing data warehousing with fuzzy technology

Abstract

A data warehouse integrates large amounts of extracted and summarized data from multiple sources for direct querying and analysis. While it provides decision makers with easy access to such historical and aggregate data, the real meaning of the data has been ignored. For example, “ Whether a total sales amount 1000 items indicates a good or bad sales performance is still unclear." From the decision makers' point of view, the semantics rather than raw numbers which convey the meaning of the data is very important. In this paper, we explore fuzzy technology to provide this semantics for the summarizations and aggregates developed in data warehousing systems. A three-layered data summarization architecture, namely, quantitative (numerical) summarization, qualitative (categorical) summarization, and quantifier summarization, is proposed. To facilitate the construction of these three summarization levels, two operators are introduced. We provide query capabilities against such enhanced data warehouses by extensions of SQL.