基于ARIMA和XGBoost算法的煤矿安全态势预测  

Prediction on Safety Situation of Coal Mine Based on ARIMA and XGBoost Algorithm

作　　者：叶黎明 施式亮 鲁义[1] 李贺 曾明圣 YE Liming;SHI Shiliang;LU Yi;LI He;ZENG Mingsheng(School of Resource&Environment and Safety Engineering,Hunan University of Science and Technology,Xiangtan Hunan 411100,China;Hunan Provincial Key Laboratory of Safe Mining Techniques of Coal Mines,Hunan University of Science and Technology,Xiangtan Hunan 411201,China)

机构地区：[1]湖南科技大学资源环境与安全工程学院,湖南湘潭411100 [2]湖南科技大学煤矿安全开采技术湖南省重点实验室,湖南湘潭411201

出　　处：《安全》2022年第2期53-59,共7页Safety & Security

基　　金：国家自然科学基金(51974120,51774135)。

摘　　要：为准确预测我国煤矿安全态势,本文提出一种基于差分自回归移动平均(ARIMA)模型和极端梯度爬升(XGBoost)算法的煤矿安全态势预测方法,该方法使用ARIMA模型对表征煤矿安全态势的3项重要指标(包括煤矿事故死亡人数、煤矿百万吨死亡率与瓦斯事故死亡人数)的历史数据进行时间序列建模,在分析单一ARIMA模型的预测结果后,使用XGBoost算法对上述3项指标的残差序列进行预测;最后,由XGBoost算法的残差预测值修正ARIMA模型预测值。结果表明:该混合模型对3项指标的预测精度均优于单一ARIMA模型,并运用此方法对2021年我国煤矿事故死亡人数、百万吨死亡率与瓦斯事故死亡人数进行预测和分析,预测方法与结果可为煤矿生产和监管部门的安全决策提供依据。In order to accurately predict the safety situation of the coal mine,in this paper,a prediction method was proposed based on the differential autoregressive moving average(ARIMA)model and the extreme gradient climb(XGBoost)algorithm.The method uses the ARIMA model to model by time series the historical data of the three important indicators,i.e.,the death toll of coal mine accidents,the fatality rate of per million tons and the death toll of gas accidents,which represent the safety situation of the coal mine.After the prediction results of single ARIMA model were analyzed,the XGBoost algorithm was used to predict the residual series of the above three indicators.Finally,the residual prediction value of the XGBoost algorithm was used to modify the ARIMA model prediction value.The results show that the accuracy of the mixed model is better than single ARIMA model in predicting the above three indicators.The method was used to predict and analyze the above three indicators in 2021 and its results can provide a theoretical basis for the safety decision-making of coal mine production and supervision department.

关 键 词：煤矿事故 安全态势 预测 ARIMA模型 XGBoost算法 

分 类 号：X928.03[环境科学与工程—安全科学]