The Collect For Data Choosing Research On Data Mining

This paper studies the process of discovering the hidden and unknown information by the advanced technology of DM through combining the use of Freight Agent Management System, which has been developed by Dalian Daxian Network Co. Ltd.This paper is made up of five parts. The first part introduces the origin of the research task, the status qua home and abroad and the sequence of the paper. The second part describes the Freight Agent Management System, giving a clear definition of the position of Decision-making sub-system in it. By explaining the data stream, control stream and all modules, readers can outline the DM. The third part introduces the DM in the area of artificial intelligence. It makes analysis of the freight agent from the point of view of the conception, the content, the essence and the core of DM. It then brings forward the methodology of SAS-SEMMA. The forth part is the main part of the paper, mainly introducing all the process of DM-data preparation, then data mining according to the SEMMA. This part puts forward the system conception of DM and the Apriori algorithm. Then it evolves the create-frequent-set algorithm which is fit for the Freight Agent Management System. Because of the shortage of efficiency, the algorithm is improved. Because some of the items are not Boolean variables, it needs the quantitative attributes association rules discovering algorithm. In general, there are the levels among the items, so multi-level association rules exist. After perfecting the algorithm, we need interpret and evaluate the knowledge. In the end, it discusses the privacy and security of DM. The fifth part relates the future problems and prospect.Through practice, DM is turned out to be useful to the freight agent department which can benefit from it.