基于SARIMA与LSTM的提前退休现状分析及预测——以天津市为例  

Analysis and Prediction of Early Retirement Status Based on SARIMA and LSTM——Taking Tianjin as an Example

作　　者：王园朝 叶炳昊 卢奕皓 陈雨露 王小超 WANG Yuanchao;YE Binghao;LU Yihao;CHEN Yulu;WANG Xiaochao(School of Mathematical Sciences,Tiangong University,Tianjin 300387,China)

机构地区：[1]天津工业大学数学科学学院,天津300387

出　　处：《数学建模及其应用》2024年第4期31-39,共9页Mathematical Modeling and Its Applications

基　　金：天津工业大学21级大学生创新创业训练计划(202310058050)。

摘　　要：随着我国人口老龄化程度不断加深,延迟退休政策的实施成为大势所趋.但现实中提前退休现象仍然严重,对经济与社会发展产生负面影响.基于此,本文以天津市为例,聚焦于职工因病或非因工伤残而提前退休的特殊现象,利用数据可视化及决策树算法探究了职工因患病而提前退休的影响因素,建立了SARIMA时间序列预测模型和LSTM神经网络模型,分别对天津市未来因患病而提前退休的职工总人数和因患某类疾病而提前退休的人数进行预测.研究发现:性别、年龄和单位类型对提前退休职工的患病情况有影响;从整体上看,在短期内天津市未来因患病而提前退休的职工人数大概率会保持稳定,有缓慢下降的趋势.研究结果旨在提出针对性的建议,为完善我国退休制度提供数据支持和理论依据.With the ageing of China's population,the delayed retirement policy has become a major trend.However,in reality,the phenomenon of early retirement is still serious,which has a negative impact on economic and social development.This paper took Tianjin as an example,focusing on the special phenomenon of early retirement due to illness or non-work-related disability,explored the influencing factors of early retirement by using data visualization and decision tree algorithm,and established the SARIMA prediction model and LSTM model to predict the total number of employees who will retire early and the number of employees who will retire early due to a certain type of illness,respectively.The study found that gender,age and type of organisation have an impact on the illnesses of early-retirement workers;overall,the number of future early-retirement workers due to illnesses in Tianjin shows an increasing trend in a short period of time,and then decreases.The results of the study are intended to make targeted recommendations to provide data support and theoretical basis for improving the retirement system in China.

关 键 词：提前退休 SARIMA时间序列 LSTM神经网络 

分 类 号：O175[理学—数学]