Research And Application Of Life Insurance Sales-oriented Business Intelligence Software

Now increasingly broad application of information systems and databases, data is animportant resource for many enterprises. How to be more rational and effective useenterprise data, is also one of the important issues facing enterprises today. The developmentof modern information technology in today’s information economy, in this new economy, theexplosive proliferation of information having a strong demand for information technology.Business intelligence is a new model of effective utilization of information technology inbusiness. The data is static, it is not subjective initiative.With the use of information systemsin the financial sector growing, financial firms have large amounts of data. In addition,according to the basic characteristics of business intelligence: topic-oriented (SubjectOriented), integration (Integrated), relatively stable (Non-Volatile), reflects the historicalchange (Time Variant), these characteristics are required by the life insurance industry anduseful.By analysis of the life insurance company, how to effectively target customers, is amajor problem facing the sales department. And how to effectively identify high-risk insuredand reduce the risk of claim is an important issue to the underwriting department.According two practical problems above, in this article, by analysis the basic businessintelligence algorithm existing, make group and improvement the KNN (K-NearestNeighbor algorithm) algorithm, neural network and rough set algorithm, to develop effectivedecision support algorithms for life insurance companies and to provide practical and effective solutions.The algorithm for sales using KNN algorithm. Add the concept of bi-directional decisionanalysis to increase the confidence level of results and practical value. Combining rough setalgorithm, to categorize the huge amounts of data to customer information, policyinformation, reducing the time complexity of the algorithm. Final sale decision algorithm forobtaining reliable and efficient.The algorithm for underwriting decision process, using Neural Network algorithm withsigmoid activation function. Through trial the actual claim rates, average claim rate tomodify the original functiom. Let the multiple steps in the process, was completed before thedecision algorithm for determining, improves the efficiency of decision algorithm. The strainand flexibility was reduced is the disadvantage. Combining rough set algorithm, tocategorize the huge amounts of data to customer information, policy information, reducingthe time complexity of the algorithm. Final sale decision algorithm for obtaining reliable,convenient and efficient.After determined the decision algorithm. Take actual testing to ensure it has a highdegree of confidence overall. And show the high confidence and low confidence, and givesrecommendations for improvement. And described the overall properties of the function.According the algorithm, analysis current mainstream software development techniquesand database technology, to select the suitable development technologies (C#.Net2010+SQL2005), and to implement the system. After development add indexes to meet therequirements and pass the system testing.