|
|
基于Keras的LSTM网络在死亡率预测的应用
|
Abstract:
随着人口老龄化程度进一步加深，加之人口死亡率的持续下降和人口预期寿命的不断延长，对死亡率有更加有效的预测，在长寿风险的研究中尤为重要。在以往的死亡率预测中，死亡率模型的时间项大多采用传统的ARIMA时间序列方法进行预测。本文选取中国人口死亡率数据，以长短期记忆(LSTM)网络和全连接层为基础，构建基于循环神经网络的LSTM学习机，用于预测Lee-Carter模型的时间项。
As the aging degree of the population deepens, coupled with the continuous decline in population mortality and prolonged life expectancy, a more effective prediction of mortality is particularly important in the study of longevity risk. In previous mortality predictions, the time terms of mortality models were mostly predicted using traditional ARIMA time series methods. This paper selects Chinese population mortality data, based on long-short term memory (LSTM) network and dense layers, constructing a LSTM learning machine based on recurrent neural network to predict the time term of the Lee-Carter model.
| [1] | Lee, R.D. and Carter, L.R. (1992) Modeling and Forecasting US Mortality. Journal of American Statistical Association, 87, 659-675. https://doi.org/10.2307/2290201 |
| [2] | Renshaw, A.E. and Haberman, S. (2006) A Cohort-Based Extension to the Lee-Carter Model for Mortality Reduction Factors. Insurance: Mathematics and Economics, 38, 556-570. https://doi.org/10.1016/j.insmatheco.2005.12.001 |
| [3] | Cairns, A.J.G., Blake, D., Dowd, K., et al. (2011) Mortality Density Forecasting: An Analysis of Six Stochastic Mortality Models. Insurance: Mathematics and Economics, 48, 355-367. https://doi.org/10.1016/j.insmatheco.2010.12.005 |
| [4] | Li, N., Lee, R. and Gerland, P. (2013) Extending the Lee-Carter Method to Model the Rotation of Age Patterns of Mortality Decline for Long-Term Projections. Demography, 50, 2037-2051.
https://doi.org/10.1007/s13524-013-0232-2 |
| [5] | Enchev, V., Kleinow, T. and Cairns, A.J.G. (2017) Multi-Population Mortality Models: Fitting, Forecasting and Comparisons. Scandinavian Actuarial Journal, 2017, 319-342. https://doi.org/10.1080/03461238.2015.1133450 |
| [6] | Li, J. and Liu, J. (2019) A Logistic Two-Population Mortality Projection Model for Modelling Mortality at Advanced Ages for Both Sexes. Scandinavian Actuarial Journal, 2019, 97-112. https://doi.org/10.1080/03461238.2018.1511464 |
| [7] | 李志生, 刘恒甲. Lee-Carter的估计和应用——基于中国数据分析[J]. 中国人口科学, 2010(3): 47-56. |
| [8] | 黄顺林, 王晓军. 加入出生年效应的死亡率预测及其在年金系数估计中的应用[J]. 统计与信息论, 2010(5): 81-86. |
| [9] | 韩猛, 王晓军. Lee-Carter模型在中国城市人口死亡率预测中的应用与改进[J]. 保险研究, 2010(10): 3-9. |
| [10] | 曾燕, 陈曦, 邓颖璐. 创新的动态人口死亡率预测及其应用[J]. 系统工程理论与实践, 2016, 36(7): 1710-1718. |
| [11] | 曹园. 基于Lee-Carter模型的我国死亡率预测[J]. 统计与决策, 2018(9): 32-36. |
| [12] | 樊毅, 张宁. 基于全人口死亡率数据的随机死亡率模型拟合效果比较[J]. 统计与决策, 2018(23): 33-37. |
| [13] | 王晓军, 陆倩. 高龄人口死亡率预测模型的比较与选择[J]. 保险研究, 2019(3): 82-102. |
| [14] | 马海飞, 肖鸿民, 赵弘宇. 一种新的死亡率模型及基于中国人口数据的比较分析[J]. 经济数学, 2020, 37(3): 99-106. |
| [15] | 肖鸿民, 马海飞, 康彦玲. 两种死亡率预测方法的比较[J]. 统计与决策, 2020, 36(23): 5-8. |
| [16] | Deprez, P., Shevchenko, P.V. and Wüthrich, M.V. (2017) Machine Learning Techniques for Mortality Modeling. European Actuarial Journal, 7, 337-352. https://doi.org/10.1007/s13385-017-0152-4 |
| [17] | Hainaut, D. (2018) A Neural-Network Analyzer for Mortality Forecasting. ASTIN Bulletin: The Journal of the IAA, 48, 481-508. https://doi.org/10.1017/asb.2017.45 |
| [18] | 陈宁. 基于神经网络死亡率模型的我国死亡率预测及应用[D]: [硕士学位论文]. 天津: 天津财经大学, 2019. |
| [19] | 张连增, 申晴, 丁宁. 自编码器在死亡率预测中的应用研究[J]. 保险研究, 2020(7): 83-93. |
| [20] | Hong, W.H., et al. (2020) Forecasting Mortality Rates Using Hybrid Lee-Carter Model, Artificial Neural Network and Random Forest. Complex & Intelligent Systems, 7, 163-189. https://doi.org/10.1007/s40747-020-00185-w |
| [21] | Giuseppe Ciaburro, Balaji Venkateswaran. R语言实现[M]. 李洪成, 译. 北京: 机械工业出版社, 2018. |
| [22] | 贾明珠. 基于机器学习的时间序列分析方法研究与应用[D]: [硕士学位论文]. 西安: 西安科技大学, 2020. |
| [23] | 杨丽, 吴雨茜, 王俊丽, 刘义理. 循环神经网络研究综述[J]. 计算机应用, 2018, 38(S2): 1-6+26. |
