Implementation Of OLAP And Data Mining Algorithm Based On Bank Housing-Loan Credit Evaluation

Data warehouse, OLAP and data mining are all the leading technologies in business intelligence area from now on. Data warehouse which integrates the system data and faces analytical data processing is the foundation of management decision. Because of the shortage of traditional report analysis, OLAP tools with powerful multidimensional functions will be chosen inevitably by enterprises. Meanwhile, data mining technology can distill the valuable information and knowledge which are hidden in historical data.With more than 5 years' construction since 1999, Dalian branch of China Construction Bank has succeeded in founding the data warehouse platform DW&MIS which integrates main management system and is on the all-around apply period. This paper is the functional extension of enterprise risk management system, which focuses on the housing-loan operation and digs rules related with credit risk. Following are the main jobs:(1) Learn the relevant knowledge and research on the credit risk of housing-loan operation according to the demand. Data related with operation is extracted from DW&MIS, with which local data warehouse and star schema are created using Analysis Services.(2) Develop OLAP software named CCB WEBOLAP based on ActiveX technology by B/S mode using PivotTable services in Analysis Services. Users can access multidimensional dataset in data warehouse by browser and can do slice, dice, drill, pivot etc.(3) Adopt several methods to pre-process operational data according to the circumstance, such as filling the blank, dealing with noise, mapping data, synthesizing fields. This step improves the data quality and found good base for data mining using decision tree algorithm next.(4) Choose DD3 decision algorithm to dig the bank housing-loan operational data, and improve the efficiency by importing support and confidence. Result shows the improvement of algorithm's efficiency via comparison. Some typical rules dug out are explained and analyzed too.The problems existed at present are summarized, and the expectation next is mentioned in the end.