Font Size: a A A

Analysis And Design Of Insurance Salesmen Fraud Pattern Recognition System Based On.NET

Posted on:2015-02-10Degree:MasterType:Thesis
Country:ChinaCandidate:J ShenFull Text:PDF
GTID:2268330428461207Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With innovative insurance products, increased competition and the development of financial markets in the face of China’s life insurance companies, external risks are more complex and diverse. How to effectively manage these risks is an important problem faced by life insurance companies. For the insurance company’s internal operational risk, the traditional way of supervision and inspection after the investigation of the cases due to the hit rate is low, check the high human and material costs, difficult recovery and other defects, can not meet the requirements of the current management of risk prevention, control and therefore the risk of pre-screening is the inevitable choice.Establish life insurance salesmen fraud pattern recognition system not only meet the pre-screening risk more effectively prevent blindness, enhance the relevance, improve the hit rate, and the system can be made to identify fraud in its infancy, so as to effectively prevent illegal behavior occurs, avoid the company suffered financial losses. Based on the reading of literature, combined with the actual business needs of insurance companies, life insurance companies for the insurance policy loan customer’s campers use this fraud, designed and developed a model system of fraud insurance salesmen.The system uses Visual Studio2005development tools, SQL Server2005Enterprise Edition database. System uses B/S mode to facilitate the management, the background using C#console development model, making more efficient running in the background.The system design technology program:do the indicators selected using analysis of variance, discriminant analysis as a pattern recognition model. After the system is set up, the340samples for inspection, including data were successfully identify fraud, fraud data identification rate of100%non-fraud sample data has12data was mistakenly identified as fraudulent, false acceptance rate of3.8%. Identify the effect of the model to meet the conditions of application scenarios and applications, greatly reducing labor costs of risk management, which greatly improve the management of risk control, and provides an effective risk control management program.The system is on the line, which can effectively identify the insurance company the insurance policy loan customers campers use this fraud, because the design is in its infancy, it is logical to identify targeted, involving only identify a class of fraud patterns, designed to serve as a catalyst for the life insurance company risk management put forward new ideas. In the ordinary course of business management, will continue to increase demand for recognition based on the type, increasing discrimination logic to enhance the insurance company’s internal risk management and control.
Keywords/Search Tags:.NET, Insurance Fraud, Discriminant Analysis
PDF Full Text Request
Related items