Font Size: a A A

Research On Decision Support System Based On Data Warehouse

Posted on:2004-11-27Degree:MasterType:Thesis
Country:ChinaCandidate:Y GeFull Text:PDF
GTID:2168360095956766Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Research on DSS started from 1980's. The architecture that includes database, model-base and rule-base was firstly proposed. Later, knowledge-base was added into the system. There are many papers about the theory of DSS, the architecture of DSS and the construction method. The DSS is divided logically into three spaces: data space, user space and rule space . With more competition, a corporation needs more and more decision support information from MIS. The traditional DSS can not meet the need of business decision . For example, we can not integrate more different sources, the interface between database and model base can't be designed easily, user interface is not very friendly, data analysis function is not enough etc.In recent years, there are some new data analysis technology such as data warehouse, on-line analytical process, data mining. In this paper, the shortages of traditional DSS are discussed. Then data warehouse technology is introduced. A new architecture of the DSS based on DW is proposed by author, added by OLAP and DM decision tree algorithm. At last, a FoodMart DSS is designed and developed by means of the new architecture and new technology. The main contents of paper are as followings:Firstly, it analyzes the new architecture of DSS based on DW. In the new architecture, the data space is substituted by DW, the user space is replaced by OLAP, the method space is added by DM. Furthermore, the connection between the three spaces is analyzed.Secondly, OLAP is used to design new user space with multidimensional data sets. Multidimensional data view is based on multidimensional data sets, which provides new analysize functions, such as slice, dice, pivot(rotate),roll-up ,drill-down and so on.Thirdly, DM technology is organized into method space. It presents the enhancement to ID3 decision tree algorithm , which combined with information gain and property probability. The algorithm is putted into useage of customer member card. The experiment shows a good enhancement which increase ten percentage . At last, the FoodMart DSS is developed with the new technology as a case for paper. The case showed that the new architecture and technology can be successful and feasible.
Keywords/Search Tags:DSS, Data warehousing, OLAP, Data mining, Decision tree
PDF Full Text Request
Related items