Abstract:In this study, several machine learning (ML) models consisting of shallow learning (SL) models (e.g., random forest (RF), K-nearest neighbor (KNN), weighted K-nearest neighbor (WKNN), support vector machine (SVM), artificial neural network (ANN), and deep learning (DL) models (e.g., long short-term memory (LSTM), gated recurrent unit (GRU), recurrent neural network (RNN), and convolutional neural network (CNN)) have been employed for predicting air pollution and its classification. The models were selected based on factors such as prediction accuracy, model generalization, model complexity, and training time. Our study focuses on analyzing and predicting the air quality index (AQI) using daily PM 10 concentration as natural pollutants and nine meteorological parameters from March 2013 to February 2022 in Zabol. We also utilized the information gain (IG) method for feature selection. Several measures including accuracy, F1 score, precision, recall, and the area under the curve (AUC), are computed to assess model performance. This study demonstrates the efficacy of DL models, particularly CNN, in predicting the AQI with remarkable accuracy. Our findings reveal that all models effectively classify air quality levels, with an AUC of 0.95 for the good class in both DL and ANN models, significantly outperforming SL models. The AUC values for the hazardous and moderate classes of DL models were also impressive, at 0.90 and 0.83, respectively, underscoring their effectiveness in critical classifications. In terms of performance, CNN achieved an accuracy of 0.60, leading the models, while RF followed closely at 0.58. RNN, GRU, ANN, and SVM each reached an accuracy of 0.57, demonstrating a competitive edge. LSTM and WKNN recorded an accuracy of 0.55, and KNN was slightly lower at 0.53. These results highlight the superior capabilities of DL models in addressing complex air quality classifications, providing invaluable insights for policymakers. By leveraging these advanced techniques, stakeholders can implement more effective strategies to combat air pollution and safeguard public health. It is worth noting that irregular monitoring of air quality data may affect the robustness of our predictions, highlighting the need for more consistent data collection to ensure an accurate representation of pollution levels.

DAerosol-NTM: applying deep learning and neural Turing machine in aerosol prediction

AIoT-driven multi-source sensor emission monitoring and forecasting using multi-source sensor integration with reduced noise series decomposition

A Self-Organizing Memory Neural Network for Aerosol Concentration Prediction

A new method for prediction of air pollution based on intelligent computation

Real-time early warning and the prediction of air pollutants for sustainable development in smart cities

Machine learning for air quality index (AQI) forecasting: shallow learning or deep learning?

Detecting Elevated Air Pollution Levels by Monitoring Web Search Queries: Deep Learning-Based Time Series Forecasting

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

Design a regional and multistep air quality forecast model based on deep learning and domain knowledge

Particulate Matter Forecasting Using Different Deep Neural Network Topologies and Wavelets for Feature Augmentation

PM2.5 Prediction Based on Random Forest, XGBoost, and Deep Learning Using Multisource Remote Sensing Data

High-Precision Microscale Particulate Matter Prediction in Diverse Environments Using a Long Short-Term Memory Neural Network and Street View Imagery

Advancing Spatiotemporal Pollutant Dispersion Forecasting with an Integrated Deep Learning Framework for Crucial Information Capture

Time Series Analysis Based Air Pollution Level Prediction Using Long Short Term Memory Recurrent Neural Network with Attention Mechanism

Long Short-Term Memory Bayesian Neural Network for Air Pollution Forecast

Data-driven predictive modeling of PM2.5 concentrations using machine learning and deep learning techniques: a case study of Delhi, India

Prediction of air pollutant concentrations based on the long short-term memory neural network

Transfer Learning with ResNet3D-101 for Global Prediction of High Aerosol Concentrations

Forecasting Smog Clouds With Deep Learning

Evaluation of Deep Learning Models for Predicting the Concentration of Air Pollutants in Urban Environments

Airborne particulate matter measurement and prediction with machine learning techniques