Data Mining For Applied Research, Economic Forecasting

With the rapid development of the computer and Internet technology, and the information and database, computers have been applied in various areas, which results in a large quantity of data. And a lot of important information is hidden in the data, while the present database system can not find the inner connection of them and can not predict the trend of the development. All these cause the phenomena of data explosion with the shortness of knowledge. Then Data Mining was born in great need. It is a process of mining different unknown but useful information and knowledge from the large-scale of incomplete, noisy, unclear or random data used in different areas.The economic system is a complicated one with continuous evolution. To get the accurate prediction of the Chinese socialist market economy, the new method of economy prediction is quite necessary. And the Self-organization Data Mining brought forward by A.G Ivaknnenko, who adopts Group Method Data Handling( GMDH) . It realizes the self-controlling of Data Mining and sets up an optimal complexity model in an objective way. This has reduced the users'intervention and has made the process of Data Mining more automatic and has made the results more objective.This thesis tells the basic theory of the Self-organization Data Mining, mainly analyses the GMDH arithmetic, which is considered to be the crucial one, and brings forward some improvement. The application of the Self-organization Data Mining and the automatic model of economy prediction has already objectively predicted and accurately analyzed the economic index and the GDP in ChengDu, which proves that the method of Self-organization Data Mining has a got good effect, has reduced the users'intervention and has improved the quality of modeling.