Air Pollution Prediction Research Articles

Spatiotemporal prediction of particulate matter concentration based on traffic and meteorological data

Air pollution threatens worldwide human health, ecosystems, and climate change. Transportation is a major contributor to air pollution. However, the link between transportation and air pollution is intricate and influenced by multiple elements. This study employs spatiotemporal causal convolutional networks to predict air pollutants by utilizing traffic and meteorological data as inputs. Using Seoul, South Korea as a case study, a dataset of 25 regional air monitoring stations is used for prediction. The results confirm that wind speed and direction significantly impact PM2.5 dispersion, while humidity positively correlates with PM2.5 concentrations and temperature shows an inverse relationship. Additionally, vehicular traffic and subway passenger numbers exhibit positive associations, attributed to automotive emissions, road dust resuspension, and heightened human activity near subway stations, respectively. The model can potentially be utilized in a real-time air pollution tracking system to facilitate prompt interventions and reduce the harmful effects of air contamination on public health.

Air pollution prediction and backcasting through a combination of mobile monitoring and historical on-road traffic emission inventories

An important challenge for studies of air pollution and health effects is the derivation of historical exposures. These generally entail some form of backcasting, which refers to a range of approaches that aim to project a current surface into the past. Accurate backcasting is conditional upon the availability of historical data for predictor variables and the ability to capture spatial and temporal trends in these variables. This study proposes a method to backcast traffic-related air pollution surfaces developed using land-use regression models by including temporal variability of traffic and emissions and trends in concentrations measured at reference stations. Nitrogen dioxide (NO2) concentrations collected in the City of Toronto using the Urban Scanner mobile platform were adjusted for historical trends captured at reference stations. The Bayesian Estimator of Abrupt change, Seasonal change, and Trend (BEAST), a powerful tool for time series decomposition, was employed to isolate seasonal variations, annual trends, and abrupt changes in NO2 at reference stations, hence decomposing the signal. Exposure surfaces were generated for a period extending from 2006 to 2020, exhibiting decreases ranging from 10 to 50 % depending on the neighborhood, with an average of 20.46 % across the city. Yearly surfaces were intersected with mobility patterns of Torontonians extracted from travel survey data for 2006 and 2016, illustrating strong spatial gradients in the evolution of NO2 over time, with larger decreases along major roads and highways and in the central core. These findings demonstrate that air pollution improvements throughout the 14 years are inhomogeneous across space.

Deep learning-based air pollution analysis on carbon monoxide in Taiwan

Global air pollution poses a threat to humanity. Specifically, CO directly affects cardiovascular and other organ tissues and leads to numerous chronic diseases and major public health problems. The effective implementation of a deep learning model for predicting variations in CO levels would enable the early formulation of policies for controlling air pollution. In this study, a seasonal gated recurrent unit (SGRU) model, which is a deep learning time-series prediction model, was developed to predict the levels of CO in Taiwan. Atmospheric CO measurements from 2005 to 2021 were collected from the Environmental Protection Administration of Taiwan and preprocessed using the Kalman filter to achieve accurate forecasting. The performance of the proposed SGRU model was compared with that of the autoregressive integrated moving average (ARIMA), seasonal ARIMA, exponential smoothing (ETS), Holt–Winters ETS, support vector regression, and seasonal long short-term memory models in terms of mean absolute percentage error (MAPE) and root mean square error. The SGRU model achieved the lowest MAPE value of 0.94, which demonstrated its superior performance. The construction of an accurate air pollution prediction model can assist government entities in formulating health and social care strategies and in planning future air pollution control measures.

ML based assessment and prediction of air pollution from satellite images during COVID-19 pandemic

ML based assessment and prediction of air pollution from satellite images during COVID-19 pandemic

Modeling of Improved Sine Trigonometric Single Valued Neutrosophic Information based Air Pollution Prediction Approach

Industrialization and urbanization air is getting polluted due to human activities. CO, NO, C6H6, etc., are the major air pollutants. The focus of air pollutants in ambient air is controlled by the climatological parameters including wind direction, atmospheric speed of wind, temperature, and humidity. Air pollution prediction is a critical sector where machine learning (ML) technique plays a major role. Its main purpose is to tackle and understand the damaging effects of air pollutants on the environment and human health. By using a range of ML techniques such as neural networks, regression, and decision trees, we could analyze historical data on air quality alongside geographical and meteorological factors. This allows us to design model that could detect patterns and predict pollution levels. By taking proactive measures such as providing timely alerts to the public, adjusting controls on emissions, and, implementing strategies to reduce pollution, we can work towards creating healthier and cleaner environments. Embracing the potential of artificial intelligence (AI) in air pollution prediction empowers us to protect the well-being of our communities and make informed decisions. Therefore, this study develops an Improved Sine Trigonometric Single Valued Neutrosophic Information based Air Pollution Prediction (ISTSVNI-APP) approach. The major objective of the ISTSVNI-APP technique is to exploit AI concepts with neutrosophic sets (NS) models for the forecasting of air pollution. To do so, the ISTSVNI-APP technique makes use of min-max normalization as the initial preprocessing step. For predicting air pollution, the ISTSVNI-APP technique uses STSVNI approach. To improve the performance of the ISTSVNI-APP technique, modified crow search algorithm (MCSA) is used for the parameter tuning of the STSVNI system. The performance evaluation of the ISTSVNI-APP method is verified utilizing benchmark dataset. The experimental outcomes stated that the ISTSVNI-APP technique gains better performance in predicting air pollution

Interpretable Machine Learning Approaches for Forecasting and Predicting Air Pollution: A Systematic Review

Interpretable Machine Learning Approaches for Forecasting and Predicting Air Pollution: A Systematic Review

Enhancing Air Pollution Monitoring and Prediction using African Vulture Optimization Algorithm with Machine Learning Model on Internet of Things Environment

An optimal solution for monitoring air pollution, the Internet of Things (IoT)-enabled system delivers real-time data and insights on the air quality within a specific location. Air pollution poses a substantial risk to human health worldwide, with pollutants like nitrogen dioxide, particulate matter, ozone, and sulfur dioxide contributing to a range of cardiovascular and respiratory ailments. Monitoring air pollution levels is critical to understand the effect on public health and the environment. Air Pollution Monitoring includes the systematic analysis and measurement of pollutant concentration in the air, through a network of monitoring stations equipped with instruments and sensors. This station provides real-time data on air quality, allowing authorities to evaluate issue warnings, and pollution levels, and implement strategies to alleviate its negative impact. Machine learning (ML) approaches are becoming more integrated into air pollution monitoring systems for enhancing efficiency and accuracy. By analyzing vast quantities of information gathered from satellite imagery, monitoring stations, and other sources, ML approaches could detect patterns, forecast pollution levels, and pinpoint sources of pollution. This study introduces Air Pollution Monitoring and Prediction using African Vulture Optimization Algorithm with Machine Learning (APMP-AVOAML) model in IoT environment. The drive of the APMP-AVOAML methodology is to recognize and classify the air quality levels in the IoT environment. In the APMP-AVOAML technique, a four stage process is encompassed. Firstly, min-max normalization is applied for scaling the input data. Secondly, a harmony search algorithm (HSA) based feature selection process is executed. Thirdly, the extreme gradient boosting (XGBoost) model is utilized for air pollution prediction. Finally, AVOA based parameter selection process is exploited for the XGBoost model. To illustrate the performance of the APMP-AVOAML algorithm, a brief experimental study is made. The resultant outcomes inferred that the APMP-AVOAML methodology has resulted in effectual outcome.

Reliable AI models can reveal key processes of heat recovery steam generator operation in air pollutant emission

Artificial intelligence (AI) algorithms have been applied to air pollution prediction modeling, but assessment of their trustworthiness is lacking. This study aims to examine whether AI-based models can identify key operational processes of a Heat Recovery Steam Generator (HRSG) facility and predict the emission of target air pollutants, including nitrogen oxides (NOx), sulfur oxides (SOx), and total suspended particles (TSP). First, key operational variables are selected based on the feature importance of the Random Forest models (that is, data-driven input selection). Secondly, Bidirectional Long Short-Term Memory-based AutoEncoder (BiLSTM-AE) and Random Forest (RF) models with the data-driven input selection are trained and evaluated using multiple predictive performance metrics. Then, the BiLSTM-AE and RF models are trained and evaluated using the operational variables selected by experts. Results demonstrate that the data-driven and expertise-based methods select five common operational variables that are key factors to predict variations of target output variables. The results from the multi-metric evaluations show a skillful prediction of BiLSTM-AE and RF models for NOx and TSP, but not SOx. For TSP prediction, RF is more sensitive to the input selection method than BiLSTM-AE. This study underscores the dependency of the prediction skill of AI algorithm-based models on the target air pollutant and input selection method, and suggests the trustworthiness of selected key operational variables by RF's feature importance.

Collaborative Learning at the Edge for Air Pollution Prediction

Collaborative Learning at the Edge for Air Pollution Prediction

Development of land use regression model to estimate particulate matter (PM10) and nitrogen dioxide (NO2) concentrations in Peninsular Malaysia

Nowadays, exposure modelling has become the preferred method for assessing human air pollution exposure due to its capability to predict air pollution under various conditions. The land use regression model (LUR) is a widely conducted model utilized to estimate air pollutants especially in unmonitored locations. However, the application of the model is still lacking in developing countries, especially in the Southeast Asia region. Therefore, this study was conducted to develop the LUR model to estimate PM10 and NO2 concentrations in Peninsular Malaysia. Multiple linear regression with a supervised forward stepwise was used to develop the models, and the models were validated using the leave-out-one cross-validation (LOOCV) approach. Results showed that the LUR model of PM10 explained 58.5% variation, while the NO2 LUR model described 86.8% variation. The difference value of PM10 model R2 and LOOCV R2 were between 0.1% and 1.2 %, and the NO2 models were between 0.01% and 0.08% depicting the robust stability of the models. Both models indicated that increased road and industrial areas significantly influence PM10 and NO2 concentrations. Nevertheless, more studies on the LUR model should be conducted in developing countries to assess the model's applicability in the region.

Two-Stage Feature Engineering to Predict Air Pollutants in Urban Areas

Two-Stage Feature Engineering to Predict Air Pollutants in Urban Areas

Spatio-temporal data prediction of multiple air pollutants in multi-cities based on 4D digraph convolutional neural network.

In response to the problem that current multi-city multi-pollutant prediction methods based on one-dimensional undirected graph neural network models cannot accurately reflect the two-dimensional spatial correlations and directedness, this study proposes a four-dimensional directed graph model that can capture the two-dimensional spatial directed information and node correlation information related to multiple factors, as well as extract temporal correlation information at different times. Firstly, A four-dimensional directed GCN model with directed information graph in two-dimensional space was established based on the geographical location of the city. Secondly, Spectral decomposition and tensor operations were then applied to the two-dimensional directed information graph to obtain the graph Fourier coefficients and graph Fourier basis. Thirdly, the graph filter of the four-dimensional directed GCN model was further improved and optimized. Finally, an LSTM network architecture was introduced to construct the four-dimensional directed GCN-LSTM model for synchronous extraction of spatio-temporal information and prediction of atmospheric pollutant concentrations. The study uses the 2020 atmospheric six-parameter data of the Taihu Lake city cluster and applies canonical correlation analysis to confirm the data's temporal, spatial, and multi-factor correlations. Through experimentation, it is verified that the proposed 4D-DGCN-LSTM model achieves a MAE reduction of 1.12%, 4.91%, 5.62%, and 11.67% compared with the 4D-DGCN, GCN-LSTM, GCN, and LSTM models, respectively, indicating the good performance of the 4D-DGCN-LSTM model in predicting multiple types of atmospheric pollutants in various cities.

Estimating PM2.5 utilizing multiple linear regression and ANN techniques

The accurate prediction of air pollutants, particularly Particulate Matter (PM), is critical to support effective and persuasive air quality management. Numerous variables influence the prediction of PM, and it's crucial to combine the most relevant input variables to ensure the most dependable predictions. This study aims to address this issue by utilizing correlation coefficients to select the most pertinent input and output variables for an air pollution model. In this work, PM2.5 concentration is estimated by employing concentrations of sulfur dioxide, nitrogen dioxide, and PM10 found in the air through the application of Artificial Neural Networks (ANNs). The proposed approach involves the comparison of three ANN models: one trained with the Levenberg–Marquardt algorithm (LM-ANN), another with the Bayesian Regularization algorithm (BR-ANN), and a third with the Scaled Conjugate Gradient algorithm (SCG-ANN). The findings revealed that the LM-ANN model outperforms the other two models and even surpasses the Multiple Linear Regression method. The LM-ANN model yields a higher R2 value of 0.8164 and a lower RMSE value of 9.5223.

Air Pollution Index Forecasting Model using Automated Machine Learning

Malaysia's Air Pollution Index (API) has steadily climbed in recent decades, posing a severe environmental hazard. In this paper, authors used the Automated Machine Learning (AutoML) technique for, to analyse daily integer value time series data for API in Malaysia from January to December in 2019. The parameters of the models will be calculated using an error rated, and their model’s result will be compared to determine which model best fits the data. The results showed that the AutoML model was more accurate in predicting API values. The MSE value of the AutoML model and the RMSE were lower, indicating that the model had a lower error rate than traditional regression models. Furthermore, predicting ability will be evaluated using mean square error (MSE) and correlation coefficient as accuracy. The findings indicated that the AutoML model produced better results in MSE and better accuracy. The fundamental benefit of utilising the AutoML model is that it is readily automated, making it a valuable tool for academics and scientists. This model may also be used as a forecasting tool to enhance air quality in a certain location. Overall, the results indicate that the AutoML model is a viable and promising technique for predicting air pollution. The findings are critical for managing API outcomes in the future and adopting protective measures for air conservation.

Machine Learning Models for Predicting Air Pollution

Abstract: The air quality monitoring system collects data of pollutants from different location to maintain optimum air quality. In the current situation, it is the critical concern,. The introduction of hazardous gases into the atmosphere from industrial sources, vehicle emissions, etc. pollutes the air. Today, the amount of air pollution in large cities has surpassed the governmentset air quality index value and reached dangerous levels. It has a significant effect on a human health. The prediction of air pollution can be done by the Machine Learning (ML) algorithms. Machine Learning (ML) combines statistics and computer science to maximize the prediction power. ML is used in order to calculate the Air Quality Index. Various sensors and an Arduino Uno microcontroller are utilized to collect the dataset. Then by using K- Nearest Neighbor (KNN) algorithm, the air quality is predicted.

Survey on Air Pollution in India: Insights, Predictions, and Mixed Gas Analysis for an Improved Future

Abstract: The sustenance of life on Earth is intricately tied to the quality of our air. But this precious resource is facing mounting threats from the harmful consequences of rapid industrialization, transportation networks, and the everyday practices of modern living. This research paper addresses the pressing issue of rising air pollution in India, with a strong emphasis on environmental well-being. Utilizing data from the Central Pollution Control Board of India (CPCB), our project focuses on air quality visualization, pollution prediction, and mixed gas analysis. We have developed an intuitive web-based platform for visualizing air quality data, offering a valuable tool for the public and policymakers. Additionally, predictive models are employed to anticipate pollution levels, facilitating timely intervention and mitigation. Our mixed gas analysis sheds light on the composition of atmospheric gases, enhancing our understanding of pollution sources. Through this interdisciplinary approach, our research aims to provide a comprehensive solution to combat air pollution, fostering environmental sustainability and public health improvement in India

Implementation of the K-Nearest Neighbor Algorithm to Predict Air Pollution

Air pollution is a serious issue that impacts air quality and human health. In this study, the K-Nearest Neighbor (KNN) algorithm is applied using Rapidminer software to predict air pollution levels. The research aims to predict air pollution levels based on various air quality parameters such as particulates, PM10, PM2.5, CO, NO2, SO2, and O3. By implementing the K-Nearest Neighbor algorithm in Rapidminer, the predicted values for air pollution data resulted in an accuracy of 93.94%. This study concludes that employing the K-Nearest Neighbor algorithm using Rapidminer software can be an effective method for predicting air pollution levels. With a strong accuracy rate of 93.94%, this can have a positive impact on both human health and the environment. The predictive model developed can aid decision-making and enhance awareness among the public regarding the importance of maintaining air quality management.

Improving PM2.5 prediction in New Delhi using a hybrid extreme learning machine coupled with snake optimization algorithm

Fine particulate matter (PM2.5) is a significant air pollutant that drives the most chronic health problems and premature mortality in big metropolitans such as Delhi. In such a context, accurate prediction of PM2.5 concentration is critical for raising public awareness, allowing sensitive populations to plan ahead, and providing governments with information for public health alerts. This study applies a novel hybridization of extreme learning machine (ELM) with a snake optimization algorithm called the ELM-SO model to forecast PM2.5 concentrations. The model has been developed on air quality inputs and meteorological parameters. Furthermore, the ELM-SO hybrid model is compared with individual machine learning models, such as Support Vector Regression (SVR), Random Forest (RF), Extreme Learning Machines (ELM), Gradient Boosting Regressor (GBR), XGBoost, and a deep learning model known as Long Short-Term Memory networks (LSTM), in forecasting PM2.5 concentrations. The study results suggested that ELM-SO exhibited the highest level of predictive performance among the five models, with a testing value of squared correlation coefficient (R2) of 0.928, and root mean square error of 30.325 µg/m3. The study's findings suggest that the ELM-SO technique is a valuable tool for accurately forecasting PM2.5 concentrations and could help advance the field of air quality forecasting. By developing state-of-the-art air pollution prediction models that incorporate ELM-SO, it may be possible to understand better and anticipate the effects of air pollution on human health and the environment.

Modeling NO2 air pollution variation during and after COVID-19-regulation using principal component analysis of satellite imagery

By implementing Principal Component Analysis (PCA) of multitemporal satellite data, this paper presents modeling solutions for air pollutant variation in three scenarios related to COVID-19 lockdown: pre, during, and after lockdown. Tropospheric NO2 satellite data from Sentinel-5P was used. Two novel PCA-models were developed: Weighted Principal Component Analysis (WPCA) and Rescaled Principal Component Analysis (RPCA). Model results were tested for goodness-of-fit to empirical NO2 data. The models were used to predict actual near-surface NO2 concentrations. Model-predicted NO2 concentrations were validated with NO2 data acquired at ground monitoring stations. Besides, meteorological bias affecting NO2 was assessed. It was found that the weather component had substantial impact on NO2 built-ups, propitiating air pollutant decrease during lockdown and increase after. WPCA and RPCA models well fitted to observed NO2. Both models accurately estimated near-surface NO2 concentrations. Modeled NO2 variation results evidenced the prolongated effect of the total lockdown (up to half a year). Model-predicted NO2 concentrations were found to highly correlate with monitoring station NO2 data collected on the ground. It is concluded that PCA is reliable in identifying and predicting air pollution variation patterns. The implementation of PCA is recommended when analyzing other pollutant gases.

Using neural networks and remote sensing for spatio-temporal prediction of air pollution during the COVID-19 pandemic.

The study aims to monitor air pollution in Iranian metropolises using remote sensing, specifically focusing on pollutants such as O3, CH4, NO2, CO2, SO2, CO, and suspended particles (aerosols) in 2001 and 2019. Sentinel 5 satellite images are utilized to prepare maps of each pollutant. The relationship between these pollutants and land surface temperature (LST) is determined using linear regression analysis. Additionally, the study estimates air pollution levels in 2040 using Markov and Cellular Automata (CA)-Markov chains. Furthermore, three neural network models, namely multilayer perceptron (MLP), radial basis function (RBF), and long short-term memory (LSTM), are employed for predicting contamination levels. The results of the research indicate an increase in pollution levels from 2010 to 2019. It is observed that temperature has a strong correlation with contamination levels (R2 = 0.87). The neural network models, particularly RBF and LSTM, demonstrate higher accuracy in predicting pollution with an R2 value of 0.90. The findings highlight the significance of managing industrial towns to minimize pollution, as these areas exhibit both high pollution levels and temperatures. So, the study emphasizes the importance of monitoring air pollution and its correlation with temperature. Remote sensing techniques and advanced prediction models can provide valuable insights for effective pollution management and decision-making processes.