Search

The interpolation, prediction, and feature analysis of fine-gained air quality are three important topics in the area of urban air computing. The solutions to these topics can provide extremely useful information to support air pollution control, and consequently generate great societal and technical impacts. Most of the existing work solves the three problems separately by different models. In this paper, we propose a general and effective approach to solve the three problems in one model called the Deep Air Learning (DAL). The main idea of DAL lies in embedding feature selection and semi-supervised learning in different layers of the deep learning network. The proposed approach utilizes the information pertaining to the unlabeled spatio-temporal data to improve the performance of the interpolation and the prediction, and performs feature selection and association analysis to reveal the main relevant features to the variation of the air quality. We evaluate our approach with extensive experiments based on real data sources obtained in Beijing, China. Experiments show that DAL is superior to the peer models from the recent literature when solving the topics of interpolation, prediction, and feature analysis of fine-gained air quality. [ABSTRACT FROM AUTHOR]

Air quality forecasting is a crucial part of megacities control and planning. It is necessary to develop high-quality forecasting methods in real-world applications, especially the issues that affect public health. Like many real-world problems, air quality is affected by uncertainties of nonstationary characteristics that their statistical attributes frequently change over time. Type-2 fuzzy systems have the capability to handle high-order uncertainties such as sequential dependencies associated with time series. This study develops a type-2 fuzzy intelligent system for air quality forecasting. Additionally, a dynamic time warping algorithm has been applied to estimate the pattern's similarity in long time series. The proposed model has been evaluated on a real dataset that contains the one-decade information about outdoor pollutants from April 2011 to November 2020 in Tehran and Beijing. The experiments have confirmed that the proposed model has lower forecasting error than counterpart models in terms of the RMSE, MAE, and MPE for both Tehran and Beijing time-series datasets. Also, the results confirm the superiority of the proposed model with an average area under the ROC curve (AUC) of 95% with a 95% confidence interval. This is promising for the prediction of air quality and make strategic prevention decisions. [ABSTRACT FROM AUTHOR]