基于灰色最优化模型的保费收入动态预测——以东北三省为例  

Dynamic Forecasting on Premium Income Using Optimal Improved Grey Model Evidence from the Three Northeast Provinces of China

作　　者：张鑫 赵苑达[1] 蒋鹏 ZHANG Xin;ZHAO Yuanda;JIANG Peng(School of Finance,Dongbei University of Finance and Economics,Dalian 116023,China;School of Economics and Management,Dalian Ocean University,Dalian 116023,China)

机构地区：[1]东北财经大学金融学院,辽宁大连116023 [2]大连海洋大学经济管理学院,辽宁大连116023

出　　处：《辽宁大学学报（哲学社会科学版）》2018年第6期46-56,共11页Journal of Liaoning University(Philosophy and Social Sciences Edition)

基　　金：国家哲学社会科学基金"保险公司过度风险承担与系统性风险跨市场传染机制研究"(17BJY204);2017年度黑龙江省哲学社会科学研究规划项目"黑龙江省农村养老保险适度水平研究"(17JYC143)

摘　　要：保费收入水平是衡量国家(区域)经济发展情况的重要指标,是准确研判市场走势并合理制定行业政策的重要参考依据,保费收入预测由于受到诸多方面因素的影响,预测精度通常不理想。基于东北三省的实证数据,运用基因演算法对经典灰色预测模型GM(1,1)中的背景值进行优化求解,并创新性地引入滚动建模以确定最优建模时点长度,构建灰色最优化预测模型ORGM(1,1),对上述地区未来的保费收入进行预测。实证研究结果表明,经创新改进后的灰色最优化模型极大地提高了预测的精度。预测结果显示2018年底东北三省保费收入将全部突破千亿大关,其中辽宁省总量最大、吉林省增速最快,未来东北保险市场将持续发力,前景光明。The level of premium income is an important index to measure the economic development of a country (region),and it is also an important reference basis to accurately judge the market trend and rationally formulate industrial policies.Due to the influence of many factors,the prediction accuracy of premium income is usually not ideal.In this paper,a novel grey forecasting model,named ORGM (1,1),is proposed.In ORGM (1,1)the genetic algorithm is used to determine the optimal background value of traditional GM (1,1). Moreover,the concept of rolling modeling is introduced innovatively to determine the optimal ordinal length of modeling.An empirical study based on the panel data of the three Northeastern provinces of China is taken out.The result of empirical study shows that the proposed ORGM (1,1)can improve the accuracy of prediction.Also,it shows that the premium income of each province will exceed 100 billion by the end of 2018,respectively;and the insurance market in Northeast of China will achieve a greater development in the future.

关 键 词：东北三省 保费收入 滚动建模 ORGM(1 1) 动态预测 

分 类 号：F840.4[经济管理—保险]