摘要
在提出和分析一套比较完整的人身保费收入指标体系基础上,采用我国1989￣2005年的实际经济年度数据,利用动量法和学习率自适应调整的BP算法对我国人身保险费收入进行拟合及测试(预测)分析。结果表明,与计量经济模型相比,神经网络模型具有更高的测试(预测)精度,平均相对误差为0.69%。最后通过相关度和影响度分析,强化了该模型应用价值。
The paper give the norm system of Personal Insurance premium income. One method of modeling the insurance based on an improved BP model is established. Fitness and prediction analysis of china Personal Insurance premium income was made using all the date of the Insurance from 1989 to 2005 .Compared with the Econometric Model. This model has higher prediction precise with relative error of 0.69%. Finally Through analyzing of correlativity and influence. BP model proved to be more effective and applicable.
出处
《价值工程》
2006年第11期96-98,共3页
Value Engineering
基金
福建省教育厅基金资助项目(JA03057S)
关键词
BP神经网络
指标体系
人身保费收入
BP neural network
norm system personal
insurance premium income
