School of economics, Changzhou University, Jiangsu, Changzhou 213159, China; Business College, Changzhou University, Jiangsu, Changzhou 213159, China.
School of Economics, Zhejiang University of Finance and Economics, Hangzhou 310018, China; Center for Regional Economy & Integrated Development, Zhejiang University of Finance & Economics, Hangzhou 310018, China.
Sci Total Environ. 2020 Dec 15;748:141428. doi: 10.1016/j.scitotenv.2020.141428. Epub 2020 Aug 2.
Foreknowledge of the air quality indicators (i.e. AQI, PM, PM, SO, CO, NO, and O) provides decision-makers a possibility for building an early-warning system and tailoring related policies and plans accordingly so as to reduce the negative influences of these pollutants. However, accurate forecasts are hardly obtained because strong seasonal variations in meteorological circumstances can largely give rise to seasonal fluctuations in the time series of these indicators, which are difficult to be described and extracted by traditional forecasting tools. To address such issues, a seasonal nonlinear grey Bernoulli model is developed to provide skillful forecasts, which can effectively grasp the nonlinear and seasonal features. Subsequently, this paper elaborates on the model and method used for parameter estimations. For validation and verification purposes, operational seasonal forecasts of the air quality indicators in the four representative cities (Shanghai, Hangzhou, Nanjing, and Hefei) in the Yangtze River Delta are performed, in comparison with five prevalent forecasting tools including SFGM(1,1), SGM(1,1), LSSVM, SARIMA, and BPNN. Results show that the proposed model outperforms other competitors in improving the prediction accuracy of the seasonal air quality changes. Thus, the verified model is recommended to produce future estimations of the air quality indicators in the Yangtze River Delta from 2020 to 2021, revealing that Shanghai, Hangzhou, and Hefei will have better air quality than before, while Nanjing will be subjected to a poorer one. Eventually, some suggestions related to the prevention of atmospheric pollution are provided to further improve air quality.
空气质量指标（即 AQI、PM、PM、SO、CO、NO 和 O）的先验知识为决策者提供了建立预警系统和相应调整相关政策和计划的可能性，从而减少这些污染物的负面影响。然而，由于气象条件的季节性变化很强，很难获得准确的预测，这些变化会导致这些指标的时间序列出现季节性波动，这很难用传统的预测工具来描述和提取。为了解决这些问题，开发了季节性非线性灰色伯努利模型来提供熟练的预测，这可以有效地掌握非线性和季节性特征。随后，本文详细阐述了模型和参数估计方法。为了验证和验证目的，在长江三角洲的四个代表性城市（上海、杭州、南京和合肥）进行了空气质量指标的季节性预测，与包括 SFGM(1,1)、SGM(1,1)、LSSVM、SARIMA 和 BPNN 在内的五种流行预测工具进行了比较。结果表明，与其他竞争对手相比，所提出的模型在提高季节性空气质量变化预测精度方面表现出色。因此，建议使用经过验证的模型对 2020 年至 2021 年长江三角洲的空气质量指标进行未来预测，结果表明上海、杭州和合肥的空气质量将比以前更好，而南京的空气质量将更差。最后，提出了一些与大气污染防治有关的建议，以进一步改善空气质量。
