Research On Life Insurance Intelligent Underwriting

In recent years,the rapid development of China's economy has led to the improvement of people's living standards,the increasing demand for old-age security,and the trend of population aging has also increased people's demand for life insurance.For insurance companies,the increase in customer demand has brought business growth,but also brought risks.Underwriting is not only the first pass for insurance companies to control customer risk,but also the most important risk defense line.At present,the insurance industry generally adopts the method of manual underwriting in the process.The increase of business volume makes the workload of underwriting and insurance personnel close to saturation.Moreover,the artificial underwriting relies too much on the subjective judgment and working level,which makes it difficult to improve the efficiency.Therefore,it is one of the important needs of insurance companies to use statistical methods to form an intelligent underwriting model and improve the quality and efficiency of underwriting.Starting from the demand of practical work and using statistical methods,this paper preliminarily constructs an intelligent underwriting model,which can discover the relevant variables of potential impact risk and calculate the conclusion of underwriting,so as to improve the efficiency of underwriting.When judging the conclusion of the intelligent underwriting model in this paper,besides considering the conclusion of the underwriting,the risk of the customer be in danger is also added.If the risk of customer be in danger is higher in the future,the final conclusion of the underwriting is not passed.Therefore,this paper is based on two aspects of risk,namely,the risk that the customer failed to pass the pre-purchase insurance link and the risk that the customer be in danger after being insured.Firstly,through data description,this paper preliminarily explores the relationship between customer group characteristics and customer risk.Based on the two aspects of risk,the paper establishes the preliminary discriminant model of insurance conclusion and the preliminary discriminant model of insurance conclusion based on logistic regression,explores the variables that have significant impact on the results of insurance by using the model,provides reference for insurance-related practitioners,and assists them in customer risk management and targeted marketing.Then,in order to improve the accuracy of the model in risk prediction,this paper establishes the optimized model of insurance conclusion based on Stochastic Forest algorithm.This paper evaluates the insured's risk from two aspects: the risk that the insured has not passed underwriting and the risk of be in danger after passed underwriting.After predicting the customer risk from two aspects in turn,we synthesize two conclusions to decide whether to underwriting the customer,that is to say,to draw the final conclusion of the underwriting.This paper combines the two risk models to form an intelligent underwriting model.In this paper,while assisting relevant practitioners to understand the characteristics of customer risk,the application of intelligent underwriting model to improve the efficiency and accuracy of underwriting is discussed effectively,and the establishment of intelligent underwriting model is preliminarily completed at a higher level of accuracy.