The Integration Design And Implementation On Data Rocessing Of Group Customers Of Commercial Bank Based ETCL

In the information era, data floods in every way of walk, including commerce, society, science, medicine and many other fields. These data gather together for some relationship and called related group. However, daily enlarged data warehouse increase the possibility o f’dirty data’, which brings many adverse factors. It causes high expenditure for warehouse maintenance, and also leads policy makers and managers to make fata misjudgment. In this thesis, ETCL, which stands for Extract, Transformation, Cleaning, and Loading, is proposed to process the large quantity of data.Based on the project of early-warning of the commercial bank customers risk, the thesis analyses the characteristic of the customers of the commercial bank that is also called related group. According to these the characteristic of large numbers, the thesis, using the software of S AS and lots of methods, inducts and extracts the data, then designs the rules of cleaning. In order to meet the modeling need, the thesis goes to a research in the data acquisition with the method of stratified sampling. The thesis designs the sampling plan, chooses the characteristic attribute and extracts the suitable samples. At last, this algorithm modeling is applied to the credit crisis early-warning system for commercial bank to realize the function of the data processing module. These data acquisition, as the base of the modeling, achieves the function of data processing. Using these powerful data the decision makers can obtain the premonitory information of the clients and the related clients group, then achieve the management before risks appear.This thesis, by the ways of SAS, realizes the ETCL and extraction of the data. It has more advantages on time and efficiency. And it has practical and applicable values. It can smooth the data processing procedure during data mining in great way.