Short-term air pollution forecasting based on environmental factors and deep learning models

Abstract

The effects of air pollution on people, the environment, and the global economy are profound -and often under-recognized. Air pollution is becoming a global problem. Urban areas have dense populations and a high concentration of emission sources: vehicles, buildings, industrial activity, waste, and wastewater. Tackling air pollution is an immediate problem in developing countries, such as North Macedonia, especially in larger urban areas. This paper exploits Recurrent Neural Network (RNN) models with Long Short-Term Memory units to predict the level of PM10 particles in the near future (+3 hours), measured with sensors deployed in different locations in the city of Skopje. Historical air quality measurements data were used to train the models. In order to capture the relation of air pollution and seasonal changes in meteorological conditions, we introduced temperature and humidity data to improve the performance. The accuracy of the models is compared to PM10 concentration forecast using an Autoregressive Integrated Moving Average (ARIMA) model. The obtained results show that specific deep learning models consistently outperform the ARIMA model, particularly when combining meteorological and air pollution historical data. The benefit of the proposed models for reliable predictions of only 0.01 MSE could facilitate preemptive actions to reduce air pollution, such as temporarily shutting main polluters, or issuing warnings so the citizens can go to a safer environment and minimize exposure.

Multi-Horizon Air Pollution Forecasting with Deep Neural Networks

“…An earlier study with a similar approach was described in [ 15 ], attempting to forecast the PM10 concentration values 3 h in the future. In [ 39 ], a similar approach was described, where a subset of the data was used to classify future values using a combination of LSTM and convolutional neural networks.…”

“…By providing different sizes of input vectors, we examined which input features contributed the best results. In [ 15 ], experiments were done with input vectors ranging from a single PM10 sensor value to a 6 feature vector. Since we were using past values to predict the feature, the input was in the form of a matrix.…”

“…The same approach can be used in cities and other areas. Taking preventive measures could decrease the influence on the air quality upfront and prevent or minimize the citizens' exposure to air with hazardous pollution [15,16].…”

Integrated Multiple Directed Attention-Based Deep Learning for Improved Air Pollution Forecasting

“…Various model-based and data-based methods have been designed to improve air pollution modeling and forecasting over the past four decades [1], [6], [7]. Conventional time-seriesbased models are among the widely utilized methods for air pollution forecasting in the literature [8]- [10]. These models comprise autoregressive integrated moving average (ARIMA) and its alternatives, like seasonal-ARIMA [11], and Holt-Winters models [12], [13].…”

“…In [29], LSTM optimized using a particle swarm optimization algorithm is applied for predicting ambient air pollutants concentrations (PM2.5, PM10 ,NO 2 , CO, O 3 , and SO 2 ). In [10], an approach combining RNN models with LSTM (RNN + LSTM) is proposed to predict PM10 particles in different places in the city Skopje. Results show that using both meteorological and air pollution measurements enhances LSTM and RNN + LSTM models' forecasting accuracy.…”