key: cord-1010500-2bl1klrh
authors: Bera, Biswajit; Bhattacharjee, Sumana; Sengupta, Nairita; Saha, Soumik
title: PM2.5 concentration prediction during COVID-19 lockdown over Kolkata metropolitan city, India using MLR and ANN models
date: 2021-08-31
journal: Environmental Challenges
DOI: 10.1016/j.envc.2021.100155
sha: 3aeb849537ba3ae625d5be54d38a1d0e72be9f40
doc_id: 1010500
cord_uid: 2bl1klrh

Kolkata is the third densely populated city of India and Kolkata stands in the World's 25 most polluted cities along with 10 worse polluted cities in India. The relevant study claims that due to the imposition of lockdown during COVID-19 pandemic, the atmospheric pollution level has been significantly reduced over the metropolitan city Kolkata like other cities of the world. The main objective of this study is to predict the concentration of PM2.5 using multiple linear regression (MLR) and artificial neural network (ANN) models and similarly, to compare the accuracy level of two models. The concentration of PM2.5 data has been obtained from state pollution control board, Govt. of West Bengal and daily meteorological data have been collected from the world weather website. The results show that non-linear artificial neural network model is more rational compared with multiple linear regression model due to its high precision and accuracy level (in respect to RMSE, MAE and R2). In this research artificial neural network (ANN) model exhibited higher accuracy during the training and testing phases (root mean square error (RMSE), mean absolute error (MAE) and R2 indicate 3.74, 1.14 and 0.91 respectively in training phase and 2.55, 4.32 and 0.69 in testing phase respectively). This model (ANN)) can be applied to predict the concentration of PM2.5 during the execution of urban air quality management plan.

The unprecedented massacre has been created through the rapid transmission and fatal aftermath of novel coronavirus in the entire world. During the end of 2019, the noxious COVID-19 has initiated to blowout its acute impact and as a consequence, an alarming nuisance has been evoked in every sphere of modern human civilization . As the outbreak of this pandemic is rapidly diffused through physical interaction, social isolation is recommended as a safest remedial measure to arrest the infectious transmission of coronavirus ( Bera et al., 2020b ; Chakraborty et al., 2020 Chakraborty et al., , 2021 . The worldwide acceptance of expanded lockdown along with social distancing proves its inevitability for weakening the terminal effects of COVID-19 ( Huang et al., 2020 ) . In India, the government strictly imposed the lockdown along with social distancing regulation from 24th March 2020 to 31st May 2020 through four phases to handle COVID-19 pandemic. Amazingly, the environmental pollution level is evidently reduced during the lockdown phase due to the stop of multi-dimensional anthropogenic actions ( Dutheil et al., 2020 ) . It has been already reported that the air pollution standard of the major metropolitan areas of India has been upgraded during the COVID-19 lockdown as a consequence of the partial pausing of different economic sectors along with developmental projects ( CPCB, 2020 ; Sharma et al., 2020 ) . In this milieu, it is stated that the tremendous threat of air pollution can trigger the probability of deadly cardiovascular and respiratory diseases ( Pope et al., 2004 ) . Suspended particulate matter (SPM) has an important role in terms of acute health disorders and environmental degradation as well as the massive concentration of SPM over an area is highly responsible to influence the regional climatic change ( Haywood and Boucher, 2000 ) .The microscopic SPM is capable to penetrate into human respiratory system and this SPM brings hazardous consequences like lethal cardiovascular and respiratory diseases ( Liu et al., 2019 ; Sahu et al., 2019 ) . Excessive concentration of PM 2.5 in lower atmosphere snatches 3.15 million lives in every year whereas globally outdoor air pollution causes 3.3million mortality per year ( Lelieveld et al., 2015 ) . It was registered that in 2015, around 27.1% deaths were caused due to chronic obstructive pulmonary disease (COPD) and the extreme accumulation of PM 2.5 was considered as a triggering factor for such disaster ( Cohen et al., 2017 must be evident that the perilous upshots of PM 2.5 and PM 10 have the potentiality to accelerate the rate of human casualties Stafoggia and Bellander, 2020 ) .

In this situation, it is stated that PM 2.5 concentration can be predicted and simulated through different scientific models such as artificial intelligence, chemical transport, linear and nonlinear regression, time series analysis etc. ( Ventura et al., 2019 ; Sun et al., 2013 ; Vlachogianni et al., 2011 ; Baker and Foley, 2011 ; Wang et al., 2012 ) . PM 2.5 has ≤ 2.5 μm aerodynamic diameters and it is regarded as the toxic component which has the competency to increase human morbidity and mortality share ( Walsh, 2014 ) . The more accurate estimation about the concentration of PM 2.5 would be predicted through the combination of few models such as adaptive neuro fuzzy inference system (ANFIS), artificial neural network (ANN), multiple linear regression (MLR), general regression neural network (GRNN). Few scholars claimed that the ANFIS model has more accuracy (concentration of PM 2.5 over Tehran and Iran) in compared to other models ( Mirzaei et al., 2019 ) . Similarly, multiple linear regression model (MLR) is also a correct and suitable method in this perspective as this model has successfully projected for the presence of NO X and PM 10 in lower atmosphere over Athens and Helsinki ( Vlachogianni et al., 2011 ) . Dust storm events in the western part of Iran and PM 2.5 concentration in Sanandaj of Iran have been predicted through MLR model considering the changing pattern of temperature over the Mediterranean Sea, Damascus Deserts and various meteorological data respectively ( Amanollahi et al., 2015 ; Ausati and Amanollahi, 2016 ) . Meanwhile, the main purpose of artificial neural network model is to detect the linear and non-linear relationship between the independent and dependent variables and this model has already effectively estimated and simulated PM 2.5 concentration over the copper mines of India. Simultaneously, this model has the efficacy to examine the air quality warning systems ( Fernando et al., 2012 ) . ANN model has also been applied to estimate the daily existence of PM 2.5 in Rio de Janeiro, Brazil ( Ventura et al., 2019 ) . The prediction accuracy of PM 2.5 concentrations largely controls public health management. Both multiple linear regression and artificial neural network models have high precession and accuracy compared with other models for short-and longterm predictions. Presently, various types of modeling methodologies (linear and non-linear) have been developed through the rapid progress of science and technology. Simultaneously, the trend of using neural networks seems to be growing in different studies by the different researcher worldwide. These classic statistical methods have been now widely used in different cases and purposes particularly in the different branches of environmental sciences like pollution modeling, environmental modeling etc. ( Perez et al., 2000 ) . ANN model was used to predict 1-hr PM 2.5 concentration in Santiago, Chile. In Kuopio, Finland and Jaipur, India this model was used for the prediction of maximum and averaged PM 10 concentration ( Chelani et al., 2002 ) . A scientific study was done over Kolkata (India) to describe the inter-dependency of particulate matters with different climatic variables within the period 2015-2017. After that random forest machine learning algorithm has been applied to predict the particulate matter concentration ( Basu and Salui, 2021 ) . Another study showed the determination of a long-term trend of different particulate matters (PM 10 & PM 2.5 ) over Kolkata metropolitan city on the basis of historical data and different statistical and deep learning algorithms ( Nath et al., 2021 ) . Ground level concentration of emitted pollutants was also measured by the Gaussian distribution model ( Bandyopadhyay, 2010 ) . Gaussian type dispersion models are widely used by the researcher, policy makers and environment planners to determine the environmental impact assessment on diverse environmental projects ( Bandyopadhyay, 2010 ) . Here, various Gaussian models are widely used in India for prediction of different pollutants. These models are required various input parameters such as meteorology, land use land cover, traffic etc. ( Aggarwal et al., 2014 ) . Recently, Gaussian model has been applied over Dhaka, Bangladesh for prediction of carbon monoxide concentration ( Ferdous and Ali, 2005 ) . Similarly, a Gaussian dispersion model has been applied over Kolkata city to predict and assessment of the concentration of one of the most important vehicular pollutants (CO) ( Majumdar et al., 2010 ) . Shahraiyni and Sodoudi (2016) analysed 36 different research works which have been carried out in different parts of the world and related with particulate matter model forecasting and concentration analysis. In these research works, around 50 percent study considered artificial neural network model (ANN) with the help of multilayer perception and feed forward back propagation network topologies method whereas around 30% studies are used multiple linear regression model to predict PM 10 concentration in urban areas. And rest studies are used different machine learning algorithms ( Dutta and Jinsart, 2020 ) . MLR and ARIMA have been successfully applied over Delhi and Hong-Kong to predict respirable suspended particulate matter ( Goyal et al., 2006 ) . ANN model was successfully applied over Delhi to predict the NO 2 dispersion ( Nagendra and Khare, 2006 ) . Multilayer perception class of ANN model has applied successfully to predict the concentration of toxic metals and PM 10 in Jaipur city ( Chelani et al., 2002 ) . In the present study, multiple linear regression model and artificial neural network model have been used for simulating and predicting the PM 2.5 concentration as a dependent variable. Meteorological data such as maximum temperature (°C), minimum temperature (°C), relative humidity (RH), air pressure (AP), wind speed (WS) etc. and gaseous factors (CO, NO 2 , O 3 , PM 10 , SO 2 ) have been considered as independent variables for better result as well as accuracy of these models. The significant fact is that the accumulation of PM 10 and PM 2.5 has been curtailed over the 22 cities of India during the lockdown in 2020 compared with the year 2017 ( Sharma et al., 2020 ) . Delhi (the capital of India) has witnessed the surprising reduction of particulate matter throughout the quarantine period ( Rodríguez-Urrego and Rodríguez-Urrego, 2020 ). Subsequently, the lockdown period improves the clarity of the entire environment and intensifies the holistic environmental restoration ( Gautam, 2020 ) . Kolkata, the economic growth pole of eastern India is enlisted among the 10 most polluted cities of India and 25 worst contaminated cities of the world ( WHO, 2011 ).This metropolitan city is brutally victimized due to the catastrophic impact of COVID-19 as almost 1261 people have lost their lives because of the menace of coronavirus as of 29thAugust 2020 ( Health and Family Welfare Department, Govt. of West Bengal, 2020 ).The previous research works have depicted that if the existence of lethal pollutants is amplified by 10 gm − 3 over the troposphere, then it may markedly upsurge the daily proportion of symptomatic novel coronavirus positive cases ( Mehmood et al., 2020 ) .The concentration of PM 2.5 over Kolkata metropolitan city has been notably dwindled about 17.5% in 2020 compared with the preceding years due to the closure of transport movement and economic actions throughout the lockdown period ( Bera et al., 2020a ) . PM is now renowned as carcinogenic to humans (IARC 2013) and it is also considered as one of the leading factors of cardiovascular diseases such as stroke, asthma, ischemic heart disease, bronchitis, chronic obstructive pulmonary disease (COPD) and estimated reduction of life expectancy (WHO 2013).Such study is carried out by high accuracy models in shortterm perspective. Long-term approach is used to find out the effects of permanent disclosure (at least ten years), while short term approach is applied to determine the severe health effects and risks, specially linked with severe air pollution or particularly of PM 2.5 .The aim of this research was therefore to fill up an existing gap of knowledge on the short-term effect of PM 2.5 on the health of the residents of Kolkata, one of the biggest agglomerations in the region of North Eastern India and to compare the results with other cities worldwide. The lockdown process was remarkably reduced the pollution rate and concentration over Kolkata and Howrah municipality area and the AQI (air quality index) had improved from poor to good category. And after analysis, it has been observed that the PM 10 & PM 2.5 are the primary sources of pollution here ( Sarkar et al., 2020 ) . A recent study has been done to analyze the spatiotemporal variation of PM 10 and NO 2 over three megacities of India (Delhi, Mumbai and Kolkata). It has been observed that a significant reduction of pollutants has been registered over the megacities during lockdown period and the concentration level of PM 10 has significantly dropped over Kolkata metropolitan area mainly due to imposed of lockdown ( Ganguly et al., 2021 ) . But the study has some limitations because the whole study has been conducted within a short time frame and in India, the air pollution parameters are not routinely updated or monitored. Different climatic parameters are the main indicator of this study. These are NO 2 , O 3 , SO 2 , maximum temperature, minimum temperature, wind speed, relative humidity etc. These parameters are selected on the basis of previous important studies ( Amanollahi and Ausati, 2019 ; Mirzaei et al., 2019 ) . This paper focuses on the simulation as well as prediction of PM 2.5 concentration with a perfect accuracy level to portray more relevant and vivid scenario in terms of the contemporary context. The main objectives of this paper are (i) to predict and simulate of PM 2.5 concentration (24th March to 31st May 2020) through the application of linear and non-linear models and (ii) to compare the accuracy level of two models by obtaining the model validation values and also detects the factors which have the maximum association with PM 2.5 concentration during the lockdown phase over Kolkata metropolitan city.

The spatiotemporal concentration data of PM 2.5 has been collected from State Pollution Control Board, Govt. of West Bengal from24th March to 31st May 2020 (lockdown period). The five automatic stations (pollution measurement) like IACS Jadavpur, Fort William, Victoria Memorial, RabindraSarobarand Rabindrabharati University have been considered to maintain the spatial integrity along with augmentation of data accuracy. The existence of PM 2.5 in the troposphere has been obtained from these stations. Other important parameters such as PM 10 , CO, NO 2 , SO 2 , and O 3 have been brought from the same monitoring stations in a daily basis and averaged it. Similarly, the daily meteorological data has been taken from the world weather website ( World Weather Online, 2020 ). The scatter plot and correlation matrix have been designed by R Studio programming software.

The regression analysis is frequently used for prediction and the objective of this model is to construct a mathematical model that can be utilized for predict the dependent variable based on the inputs of independent variables or the predictors ( Juneng et al., 2011 ) . MLR model has been used to obtain the significant relationship as well as correlation between the dependent variable and the predictors or the independent variables ( Table 1 ) . Here, around 75% data has been used for training and 25% data has been applied for testing of this model. Statistical Package for Social Science (SPSS version 25) has been used here to run the model accurately.

Multiple Linear Regression model follows:

Where B 0 refers for Y-intercept; whereas X1, X2…..X n stands for the independent variables and B 1 , B2…B n are the coefficients of independent variables and refers for the error term; and Y is the dependent variable. Fig. 2 . The importance of independent variables in terms of the prediction of PM 2.5 accumulation.

Co-linearity happens when two models have a linear relationship. It creates the individual contribution of each variable and introduces redundancy. Similarly, it makes the model excessively sensitive to the data. Here, multi-collinearity problem is varied by Variance of Inflation Factor (VIF) and the value of VIF while it is less than 10 which indicates that there is no multi-collinearity problem and the regression is fit.

The VIF follows

Where, VIF stands for variance inflation factor, R 2 mean multiple coefficient of determination in a regression.

Artificial neural network (ANN) model was primarily suggested by McCulloch and Pitts who were enthused by neural network systems and the brain of living organisms. The artificial neural network model is a basically statistical and mathematical based complex interconnection which characterizes the biological neurons that are fundamental for the human brain processes. The multi-layer perceptron (MLP) is universally applied in ANN model. ANN has three diverse layers such as -i) input layer in which the data are distributed over the network, ii) hidden layer where the data are processed and finally iii) the output layer where the results for certain inputs are taken out ( Amanollahi and Ausati, 2019 ) ( Fig. 1 ). The ANN model is designed through SPSS (v25) software. There are more than one hidden layer and a key parameter. The ANN model has been broadly and commonly used for air quality estimation, simulation and prediction purposes ( Alimissis et al., 2018 ) . Here, around 75% data is used for training and 25% data has been applied for testing of the model. In ANN, a hyperbolic tangent or a sigmoid function has been considered for mathematical expediency.

Hyperbolic tangent follows,

The derivative of loss function is designed by the gradient descent method which is associated to back propagation. The following equations are maintained the algorithm of back propagation,

The square error function is

Where, E is the loss of the output y and target value t. t denotes output of training samples. y is the output of the neuron. For each neuron j and output is defined as

Where, the activation function is non-linear. The main activation function is logistic function which is followed as,

The input ( ) to a neuron is the weighted sum of outputs of previous neuron.

Any scientific prediction models require validation for determination of its performance over the dependent variable. There are three different methods which have been used here for the validation of the models. These are the root mean square error (RMSE), mean absolute error (MAE) and Pearson's correlation coefficient (R). Mean absolute error (MAE) and Root mean square error (RMSE) show the average error of the models. Both techniques are ranged from 0 to oo and both are the negative oriented values or scores. So the lower value of RMSE and MAE shows better results in the model prediction. Whereas, R 2 is the statistical parameter which has been used for the validation of the model and it ranges from 0 to 1. The R 2 value near 1 specifies the strong association between variables and contraries the lowest association between variables. Standardized coefficient and the changing pattern of R 2 are used to determine the importance of a factor in any regression model. Here, on the basis of these values the importance of these predictors is determined. The models are not highly fitted here due to some data shortage and completed with short time frame.

These equations are given below,

Where, ( − ̄ ) signifies differences and N denotes sample size.

| − ̄ | shows the absolute errors, N points out the number of errors.

Where, denotes values of x variable in the sample. ̄ specifies mean value of x variable. stands for the value of y variable and ̄ means the value of y variable.

In the current study, the result showed that multiple linear regression (MLR) is the high-quality model which is used here for the prediction of the PM 2.5 concentration over Kolkata during the lockdown. The coefficients of different corresponding predictors are mentioned in Table 1 . The result of different validation methods in training phase of MLR modelshows the model validation and the summarized result of this section is presented here, RMSE = 3.77, MAE = 1.69 and R 2 = 0.833. On the other side, the result of testing phase section is stated here, RMSE = 3.33, MAE = 5.19 and R 2 = 0.0.510. A better R 2 result for prediction and compared to the regression model usually might be due to the uneven distribution of sample data ( Zhao et al., 2018 ) .

Artificial neural network model is another type of model which is used here for simulation and prediction of PM 2.5 concentration over Kolkata ( Fig. 1 ) . In this case, an input layer, hidden layer and output layer have been incorporated for prediction. The ANN model has widely applied by the researchers for quantification and air quality prediction ( Suleiman et al., 2019 ; Radojevic et al., 2019 ) . The concise result of the applied ANN model in training stage has represented such as RMSE = 3.74, MAE = 1.14 and R 2 = 0.916 while the result of testing phase section highlights the RMSE = 2.55, MAE = 4.32 and R 2 = 0.697. It has been noticed that PM 10 is the most efficient predictor of PM 2.5 with high importance value whereas O 3 is the least effectual predictor over PM 2.5 with a low level of importance value ( Table 2 ; Fig. 2 ).

In this research two different machine learning algorithms have been applied for prediction of PM 2.5 concentration during the lockdown period (24th March to 31st May). From the comparison between the results of the two models it has been represented that the ANN model has better prediction precision and simulation result due to its lower RMSE ( Fig. 3 ) . Here, the average concentration of PM 2.5 of each day during the study period has been calculated. The RMSE and MAE value of ANN model is 3.74 and 1.14 respectively in the training phase. Whereas the RMSE and MAE value of MLR model is 3.77 and 1.69 respectively in training phase. In the case of testing phase the RMSE and MAE value is 2.55 and 4.32 respectively in the case of ANN model whereas RMSE and MAE value is 3.33 and 5.19 respectively in the case of MLR model. The prediction ofPM 2.5 concentration is a complicated issue because it can be easily affected by different factors or the predictors. The correlation (based on Pearson's method) matrix shows the degree of association between variables ( Table 3 ; Fig. 4 ). It has been exhibited that PM 10 , CO, SO 2 and NO 2 are highly correlated with PM 2.5 that means the air pollution level mostly controls on the concentration of PM 2.5 ( Table 2 ; Table 3 ) whereas the meteorological factors (wind speed and relative humidity) is negatively inter-related with the concentration of PM 2.5 ( Fig. 5 ) . This scenario showed that rainfall and high wind speed are competent to get back the atmospheric purity and directly diminish the PM 2.5 concentration over the lower atmosphere. Comparatively, the better results of these models can be derived during the cold season compared with the warm season ( Mirzaei et al., 2019 ) . It is supposed that the temperature inversion during the cold season may take a vital role because this event restricts the suspended particulate matters in the lower part of the atmosphere and subsequently decreases the accuracy level of the models.

As Kolkata metropolitan city is labelled as one of the worst polluted cities in India as well as in the world, the future projection of its atmospheric fatal pollutants would be beneficial for the human health safety as well as environmental cleanliness. This research highlights the comparison between the precision of linear model i.e., MLR and nonlinear model i.e., ANN in the aspect of prediction the occurrence of PM 2.5 in Kolkata amidst the lockdown period. The entire study signifies that the nonlinear model has exhibited the more precise prediction of PM 2.5 accumulation over this metropolitan city compared with the linear model. The comparative analysis between the two above-mentioned models focuses that the ANN model has attained the maximum perfection in case of training and testing stages for predicting the existence of PM 2.5. The most appropriate model is ANN and it is principally composed of three distinct layers. So, it must be concluded that Artificial Neural Network (ANN) has designed to predict the concentration of PM 2.5 over the worst polluted city Kolkata amid the lockdown session compared with Multiple Linear Regression (MLR) model. This artificial neural network model is very rational model which can apply to estimate the spatiotemporal concentration of PM 2.5 over any city's of the world during the implementation of long term environmental management plan.

Subsequently, the application of lockdown system is not a permanent solution to combat the threat of pollution. So, a substitute sustainable management method should be applied to maintain the cleanliness of environment. Relevant applied researches focused that plants are the primary receiver of various types of air pollutants and perform as a massive sink ( Kaur and Nagpal 2017 ; Letter and Jager, 2020 ) . A contemporary study revealed that the important plant species have high absorption capability for definite pollutants ( Salih et al., 2017 ;  Table 4 ). Introduction of air pollution-tolerant species in urban vacant spaces is highly necessary to improve the environmental health along with ecosystem values of urban life ( Bamniya et al., 2011 ) . The expansion of green encircle by plantation of tolerant species can definitely reduce the high air pollution to a certain level.

Air pollution modelling -a review

Spatial estimation of urban air pollution with the use of artificial neural network models

PM 2.5 concentration forecasting using ANFIS

Effect of the temperature variation between Mediterranean Sea and Syrian deserts on the dust storm occurrence in the western half of Iran

Assessing the accuracy of ANFIS, EEMDGRNN, PCR, and MLR models in predicting PM 2.5

A nonlinear regression model estimating single source concentrations of primary and secondarily formed PM 2.5

Harmful effect of air pollution on physiological activities of Pongamia pinnata (L.) Pierre

Dispersion modeling in assessing air quality of industrial projects under Indian regulatory regime

Estimating Particulate Matter Concentrations from MODIS AOD Considering Meteorological Parameters Using Random Forest Algorithm. Spatial Modeling and Assessment of Environmental Contaminants. Environmental Challenges and Solutions

Significant impacts of COVID19 lockdown on urban air pollution in Kolkata (India) and amelioration of environmental health

Human behavior, trustworthiness, and attitude during COVID-19 lockdown in Indian modern societal and cultural antiquity

Cleaning the river Damodar (India): impact of COVID -19 lockdown on water quality and future rejuvenation strategies

Eco-restoration of river water qualityduring COVID-19 lockdown in the industrial belt of eastern India

Prediction of ambient PM10 and toxic metals using artificial neural networks

Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study

Impact of lockdown (25th March to 15th April) on air quality. Ministry Environ. Forest Clim. Change, Govt. of India, Delhi 1-62

COVID-19 as a factor influencing air pollution? Environ Pollut

Risks to health from ambient particulate matter (PM2.5) to the residents of Guwahati city, India: an analysis of prediction model

Air Quality Modelling for Predicting Traffic Pollution in Dhaka City

Forecasting PM 10 in metropolitan areas: efficacy of neural networks

Short-term impacts of air pollutants in three megacities of India during COVID-19 lockdown

COVID-19: air pollution remains low as people stay at home. Air Qual. Atmosphere Health

Statistical models for the prediction of respirable suspended particulate matter in urban cities

Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: a review

Enhanced Secondary Pollution Offset Reduction of Primary Emissions during COVID-19 Lockdown in China

Factors influencing the variations of PM10 aerosol dust in Klang Valley, Malaysia during the summer

Evaluation of air pollution tolerance index and anticipated performance index of plants and their application in development of green space along the urban areas

The contribution of outdoor air pollution sources to pre-mature mortality on a global scale

Simulating the potential of trees to reduce particulate matter pollution in urban areas throughout the year

Evaluation of particulate matter deposition in the human respiratory tract during winter in Nanjing using size and chemically resolved ambient measurements

Assessment of vehicular pollution in Kolkata, India, using CALINE 4 model

Can PM 2.5 pollution worsen the death rate due to COVID-19 in India and Pakistan?

Evaluation of linear, nonlinear, and hybrid models for predicting PM2.5 based on a GTWR model and MODIS AOD data

Artificial neural network approach for modelling nitrogen dioxide dispersion from vehicular exhaust emissions

Long-term time-series pollution forecast using statistical and deep learning methods

Prediction of PM2.5 concentrations several hours in advance

Cardiovascular mortality and long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of disease

Potential gain in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China.: a modelling study based on nationwide data

The significance of periodic parameters for ANN modelling of daily SO 2 and NO X concentrations: a case study of Belgrade, Serbia

Air quality during the COVID-19: PM 2.5 analysis in the 50 most polluted capital cities in the world

Health risk associated with potential source regions of PM2.5 in Indian cities

Use of some trees to mitigate air and soil pollution around oil refinery, Kingdom of Bahrain wastewater

Assessing the immediate impact of COVID-19 lockdown on the air quality of Kolkata and Howrah

Statistical Modeling Approaches for PM10 Prediction in Urban Areas; A Review of 21st-Century Studies

Effect of restricted emissions during COVID-19 on air quality in India

Short-term effects of air pollutants on daily mortality in the Stockholm county -a spatiotemporal analysis

Applying machine learning methods in managing urban concentrations of traffic-related particulate matter

Prediction of 24-hour-average PM 2.5 concentrations using a hidden Markov model with different emission distributions in northern California

Forecast of daily PM2.5 concentrations applying artificial neural networks and Holt-Winters models

Evaluation of a multiple regression model for the forecasting of the concentrations of NOx and PM 10 in Athens and Helsinki

PM2.5: global progress in controlling the motor vehicle contribution

A novel coronavirus outbreak of global health concern

Application of residual modification approach in seasonal ARIMA for electricity demand forecasting: a case study of China

Short period PM2.5 prediction based on multivariate linear regression model

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.