时间尺度对时间序列模型预测城市生活垃圾清运量的影响

The Influence of Time Scale on the Prediction of the Removal Transport Weight of MSW by Time Series Model

摘要: 精准预测城市生活垃圾(MSW)的清运量有助于对其进行有效地减量化处理，数据的时间尺度是影响预测精度的关键因素。本文将全国和北京市的MSW清运量划分为1991~2022年(32年)、2000~2022年(23年)和2010~2022年(13年)三个时间尺度的数据，分别采用时间序列模型中的趋势外推模型、修正趋势外推模型、ARIMA模型和平滑ARIMA模型对其进行预测研究。综合对比发现，采用平滑ARIMA模型对2000~2022年两个空间尺度下的MSW清运量进行预测的拟合效果最好，预测精度最高，分别为99.54%和99.07%。对2023~2030年的清运量数据预测发现，预计到2030年全国MSW清运量将会上涨26.38%，北京市MSW清运量将会上涨25.36%。研究结果反映了不同时间尺度对时间序列模型预测MSW清运量精度的影响，对时间序列模型的预测研究有理论意义，可为MSW减量政策的制定提供数据参考。

Abstract: Accurate prediction of the removal transport weight of municipal solid waste (MSW) is helpful to carry out effective reduction. Data time scale is the key factor affecting the prediction accuracy. This paper divides the removal transport weight of MSW in China and Beijing into three time scales: 1991~2022 (32 years), 2000~2022 (23 years) and 2010~2022 (13 years). The trend extrapolation model, modified trend extrapolation model, ARIMA model and smooth ARIMA model are used to predict the time series model. It is found that the smooth ARIMA model has the best fitting effect and the highest prediction accuracy of 99.54% and 99.07% respectively for the removal transport weight of MSW in 2000~2022 under two spatial scales. The forecast of the removal transport weight of MSW from 2023 to 2030 finds that the removal transport weight of MSW in China is expected to increase by 26.38% by 2030, and the removal transport weight of MSW in Beijing will increase by 25.36%. The research results reflect the influence of different time scales on the accuracy of the removal transport weight of MSW prediction by time series model, which has theoretical signifi-cance for the prediction research of time series model and can provide data reference for the for-mulation of MSW reduction policy.