煤矿百万吨死亡率预测方法对比研究

Comparative Study of Prediction Methods of Death Rate Per Million-ton in Coal Mine

为了能更好地掌握煤矿安全生产状况及其发展规律,本文分别运用灰色系统理论GM(1,1)模型、残差GM(1,1)模型、Holt线性趋势模型和ARIMA(0,2,1)模型对我国2000年至2017年的煤矿百万吨死亡率数据进行分析建模,比较了四种预测方法的误差大小。通过模型的拟合结果与实际结果的对比发现,灰色系统理论的残差GM(1,1)模型预测的相对误差较小,精度较高。运用该模型对我国2018年和2019年的煤矿百万吨死亡率进行了预测,结果为0.048和0.02。该结果对我国制定煤矿安全生产宏观政策具有一定的指导意义。

In order to grasp the safety production conditions of coal mines and their development laws better,this paper uses the general GM（ 1,1） model and residual GM（ 1,1） model of grey system theory,and the Holt linear trend model and ARIMA（ 0,2,1） model based on SPSS time series analysis to analyze and predict the date of death rate per million-ton in coal mine from 2000 to 2017 in China. Through the comparison of the model＇s fitting results with the actual results,the residual GM（ 1,1） model has relatively small relative errors and high precision. The prediction of death rate per million-ton in coal mine in 2018 and 2019 is 0. 048 and0. 02,and it will provide an effective theoretical basis for the safety management of coal mines.