With the development of the insurance industry, insurance fraud is increasing rapidly. The existence of insurance fraud considerably hinders the development of the insurance industry. Fraud identification has become the most important part of insurance fraud research. In this paper, an improved adaptive genetic algorithm (NAGA) combined with a BP neural network (BP neural network) is proposed to optimize the initial weight of BP neural networks to overcome their shortcomings, such as ease of falling into local minima, slow convergence rates and sample dependence. Finally, the historical automobile insurance claim data of an insurance company are taken as a sample. The NAGA-BP neural network model was used for simulation and prediction. The empirical results show that the improved genetic algorithm is more advanced than the traditional genetic algorithm in terms of convergence speed and prediction accuracy.

随着保险业的发展，保险欺诈现象迅速增加。保险欺诈的存在极大地阻碍了保险业的发展。欺诈识别已成为保险欺诈研究中最重要的部分。本文提出了一种改进的自适应遗传算法（NAGA）与BP神经网络（BP神经网络）相结合，以优化BP神经网络的初始权重以克服其缺点，例如容易陷入局部极小，收敛缓慢率和样本依赖性。最后，以一家保险公司的历史汽车保险索赔数据为样本。将NAGA-BP神经网络模型用于仿真和预测。