|
After years of high speed development in computer and software technology, a huge amount of data is producted and stored in database, but kept collect data is the result of organization becomes data rich and little knowledge.We collected so many data and the data stored on the actual use began limited. The insurance company history business data included a huge amount of valuable information, however these information are not being used as well as they arise, so analysis these information is very important, but more importantly, through the information we can get customer risk expectations.And then,we can grade the customer risks, the low risk customer can enjoy different grades and personalized discounts and service, so as to keep customers and develop new insurance customer, in the expanding market share while increasing customer loyalty. Only in this way,we can improve the value of business data and enhance the comprehensive competitiveness of insurance company.The main job of this paper is as follows:First,Introduce the basic principle, application range and the insufficiency of data mining algorithm which based on SQL Server 2008 platform, and author of the modification program.Second,In the management software platform of SQL Server 2008, using data mining technology to analysis and obtain feature extraction in data of hunan changsha pice, since 2004, more than 200 million policies and owners, test and evaluation model, in order to get the best customer risk forecasting model.Third,Evaluate the customer risk by improvement data mining model which established through data mining technology with weighted by customer history yearly risk rate.Fourth,Make the customer risk assessment model applied in the insurance company of hunan pice sales, manage compensate and service process.Finally, This paper analyzed the feasibility of research achievements and possible problems, for possible problems, the paper provides the improvement suggestions on further research focus
|