AUTHOR=Yin Ziqi , Fang Xin 

TITLE=An Outlier-Robust Point and Interval Forecasting System for Daily PM2.5 Concentration

JOURNAL=Frontiers in Environmental Science

VOLUME=Volume 9 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2021.747101

DOI=10.3389/fenvs.2021.747101

ISSN=2296-665X

ABSTRACT=Air pollution forecasting, typified by PM2.5, can not only present effective warning information for public, but also provide decision support for control and treatment of air pollution problems. However, there are still some challenging issues that need to be solved urgently, such as outlier handling and modeling, lower forecasting stability, correction of forecasting results, and so on. In this context, this study proposes an outlier robust forecasting system to attempt to tackle above mentioned issues and bridge the research gap in current studies. Specifically, the developed system consists of two parts, i.e., point and interval forecasting parts. For point forecasting, data preprocessing module is proposed based on outlier handling and data decomposition to mitigate the negative influence of outlier and noise, which can also help the model capture the main characteristic of original time series. Meanwhile, an outlier robust forecasting module is designed for better modeling the preprocessed data. For the model to further improve accuracy, the forecasting results correction module based on error ensemble strategy is developed, which can provide more accurate forecasting results. Finally, interval forecasting part is conducted based on the newly proposed artificial intelligence-based distribution and point forecasting results to present the range of future changes. Experimental results and analysis utilizing daily PM2.5 concentration from two capital cities of China are discussed to verify the superiority and effectiveness of the developed system, which can be considered as effective technique in point and interval forecasting for daily PM2.5 concentration.