Research And Application Of The Method To Bulid Data Warehouse Based On Linguisitc Representation

Data warehouse is one of the most active branches of database studying, developing and application, and also the key factors of DSS (Decision Support System). In recent years, the progress made in data warehouse technology accelerates the improvement of DSS, which is based on data warehouse technology. At the mean while, impulse by the improvement of DSS, OLAP (On-Line Analytical Processing) technology is progressing continuously. On the other hand, in order to get over the shortcoming of classical set theory, under which each member is required to belong to a unique set, the fuzzy set theory extends the classical set theory, providing a status between in-set and not-in-set as the smooth transferring status. Because of the development of fuzzy mathematics in recent years, fuzzy mathematics has been applied in many fields such as linguistic representation, simulation technology and multimedia identification.In order to enlarge the application scope of DSS, which is based on the data warehouse technology, the function of fuzzy mathematics on linguistic representation is applied in building the data warehouse. Basing on the traditional data warehouse building method, this paper improves on the data processing and expounds the method of building the data warehouse based on linguistic representation. According to the processing subject and processing mode, this method classifies data processing in three different levels: quantitative (numerical) summarization, qualitative (categorical) summarization, and quantifier summarization. Basing on the three levels'data processing, a new way to manage the fuzzy data is put forward. In this mentioned way, which can be used for reference in future linguistic application in data warehouse, a query column is added to realize the function of fuzzy OLAP and fuzzy query.Taking the design and development of Beijing Tobacco DSS as the background, the proposed method is applied in Beijing Tobacco DSS implementation. As proved by the test and the running-in, DSS based on linguistic representation, which support fuzzy OLAP and fuzzy query, get the satisfaction of the users by leading them to deeper and more extensive data analysis.