自回归滑动平均模型法预测宁波市奉化区的人口出生率变化趋势

Trend prediction of birth rate in Fenghua district of Ningbo by autoregressive integrated moving average model

目的探讨利用自回归滑动平均模型法(ARIMA)对宁波市奉化区人口出生率数据进行预测的可行性。方法利用R3.5.0软件对浙江省宁波市奉化区1983—2013年出生率数据拟合ARIMA模型,对模型参数与残差进行统计学分析,并利用拟合的模型对奉化区2014—2018年的出生率进行预测。结果拟合的模型为ARIMA(0,1,0),模型的残差分析表明残差符合白噪声过程。2014—2018年出生率预测相对误差率最大的年份是2017年为23.40%,相对误差率最小的是2015年为3.25%。结论ARIMA(0,1,0)模型能较好地拟合奉化区出生率的时间变化趋势,可用于未来奉化区出生率的短期预测。

Objective To explore the feasibility of using the autoregressive integrated moving average model(ARIMA)to predict the birth rate of Fenghua,Ningbo.Methods R3.5.0 software was used to construct the ARIMA with birth rate data from 1983 to 2013 in Fenghua of Ningbo,and parameters and residual error were statistically analyzed.Then,this ARIMA was used to predict the birth rate from 2014 to 2018 in Fenghua.Results The optimal model was ARIMA(0,1,0),and the residual error was in accord with the white noise process.Prediction results from 2014 to 2018 showed that the maximum relative error rate of birth rate was 23.40%in 2017,and the minimum relative error rate was 3.25%in 2015.Conclusions ARIMA(0,1,0)model can be well in accord with the trend of the birth rate in the Fenghua district.We can use it for the short-term prediction of the birth rate in the Fenghua district in the future.