Evaluation Models for Choosing Insurance Policy Using Neural Networks

In this paper, we use feed forward neural networks with the back-propagation algorithm to build decision models for five insurances including life, annuity, health, accident, and investment-oriented insurances. Six features were selected for the inputs of the neural networks including age, sex, annual income, educational level, occupation, and risk preference. Three hundred insurants from an insurance company in Taiwan were used as examples for establishing the decision models. There experiments were conducted in this study. The first one considered the five insurances as a whole and established a single neural network integrating all of the five insurances in its structure. The second one built five individual neural networks independently for the five insurances, respectively. The first two experiments used the purchase records of primary and additional insurances as experimental data. The last experiment only used the data of primary insurances. We showed the experimental results and discussed problematic issues on the experiments. Based on the experimental results, the suggestions for build classifiers for insurance purchase are drawn as follows: (1) using five independent neural networks to classify the five insurances independently is better than using one single neural network to classify the five insurances simultaneously; (2) using the data of primary insurances purchases is better than using the data of primary and additional insurance purchases. Finally, the possible directions for future studies are provided at the end of this paper. Key-words: Insurance policy; Decision making; Neural Networks, Back-propagation algorithm; Classification