Research And Design Of Tax Revenue Analysis Prediction System Based On The Data Warehouse Theory

With the rapid development of Data Warehouse, OLAP and Data Mining technology, great progress has been achieved in Decision Support System. An integrated DSS is mainly based on a Data Warehouse platform, with the use of OLAP and Data mining tools. It is indispensable component and closely related to economic development to research analysis and predicted system based on data warehouse for taxation.The thesis is mainly concerned with the design and realization of the analysis and prediction system based on Data Warehouse in the field of local taxation after explaining relative theoretical knowledge started with an actual research project. In this paper, the subject background, the development of computer in taxation and present research and development trend at home and abroad are firstly produced. Secondly, the primary concepts, data construction and the design method and step of DW are narrated. Then the technology of On-line Analytical Processing (OLAP) and Data Mining (DM) in DW tools layer is illustrated in detail. We take the collection as example to illustrate application theory and realization method of data warehouse in practical tax system. Next, the solution of the analysis and prediction system is put forward; all parts of function are analyzed, and applied OLAP processing technology and Data Mining to set up the analysis system. Finally, the revenue prediction is summarized and some developing directions of DW in the future are predicted.The emphasis of this paper is on the methods, model and steps of creating data warehouse, and on the discussion of implementation of the algorithms. Moreover, OLAP database is produced through constructing the data converge aimed at collection. How to implement the several revenue prediction algorithms included interpolation polynomial, line fitting and LN function fitting is discussed based on the structure of the analysis and prediction system which is designed by applying OLAP and data mining technology. For the sake of obtaining the data which is best close to real data, the prediction model is improved through mining the laws of taxation fluctuation. The forecast algorithm is based on the best result from this model, which can increase the precision of revenue prediction greatly and provide good basis for managerial decision.