Based On Data Warehouse Technology Research And Application Of The Life Insurance

With the deepening of institutional reform in the insurance industry, the insurance market was presenting diversified competition tendency. And at the same time, since joining the WTO, many foreign insurance enterprises began to pour into China; it impacted domestic insurance market seriously. So it became a common problem for the domestic insurance enterprises that how to deal with the trapped situation; In order to improve the market competitiveness, many companies have stepped up the intensity of the enterprise informational construction; Data warehouse whose target is the business intelligence (BI) will bring the enterprise information construction to a new strategic perspective undoubtedly, it can improve the transfer efficiency of the enterprise’s internal business data effectively, and guarantee the decision of the enterprise management by constructing a global data views.Based on an actual construction project with data warehouse in a domestic life insurance enterprise, the paper has carried on a deep research on how to build the enterprise-level data warehouse and application about life insurance. Firstly, we inquired into the current research about data warehouse both at home and abroad systematically, and introduced the research content and significance. Secondly, on basis of the life insurance business, we realized the subject domain analysis of data warehouse about the life insurance, and put forward a kind of universal multiplayer data warehouse architecture system in life insurance industry. And then by contrasting the implementation technology of data warehouse, Formed a relatively complete implementation scheme of data warehouse. The related results of the paper have been used in the data warehouse system of this life insurance enterprise, at present, it works well.After that, based on the existing life insurance data warehouse platform, we chose the C5.0algorithm of decision trees and tree augmented Bayesian Network classification algorithm to build customer churn prediction model by a comprehensive comparison of customer churn modeling technology and factors, and then we began to undertake comparative assessment about these models with the index of accuracy and gain ratio, and clarified the commercial application of the model further.