Air Quality Index prediction using an effective hybrid deep learning model

A significant challenge in air quality index (AQI) prediction is to accurately evaluate the potential outcomes after conducting interventions in pollutant factors such as industrial emissions for each enterprise. Existed methods often suffer from spurious correlations caused by unmeasured confounders and are lack of interpretability of the model, leading to sub-optimal prediction performance. This motivates us to propose a causal AQI prediction framework (CAP) that employs a structural causal model (SCM) to characterize the causal structural variability of various AQI factors for robust AQI prediction. Specifically, we employ the front-door adjustment to explicitly eliminate unmeasured confounders by intervening in industrial emissions from the target enterprise. Meanwhile, we take industrial emissions of neighboring enterprises into account when intervening in the target enterprise and simulate the dispersion of industrial emissions through a Gaussian plume model based on meteorological factors. Experiments on two real-world datasets validate the superior performance of our model on AQI prediction compared to the state-of-the-art baselines.

Air pollution remains a major global concern due to its significant impact on public health and environmental sustainability. This study aims to develop a reliable global Air Quality Index (AQI) prediction model by evaluating five regression-based machine learning algorithms, including Linear Regression, Support Vector Regression, Random Forest, XGBoost, and LightGBM. The dataset contains over twenty thousand pollutant concentration records from multiple countries. Since the dataset consists of independent pollutant observations without timestamps or temporal sequences, this research employs supervised regression techniques rather than time-series forecasting methods to ensure methodological consistency with the non-temporal structure of the data. The methodology includes data preprocessing, validation of geocoded country information for missing values, transformations to address skewed pollutant distributions, and feature selection based on established environmental standards. Sample weights were applied to account for uneven regional representation, and systematic hyperparameter tuning with cross-validation was conducted to optimize model parameters and reduce potential overfitting. Evaluation metrics are supported by correlation analysis to quantify relationships between pollutants and AQI. The results show that XGBoost delivers the highest and most stable performance, with a MAE of 0.0216, MSE of 0.0010, RMSE of 0.0318, R² of 0.9971, and MAPE of 0.5664. Feature importance analysis highlights PM2.5 as the most influential pollutant, followed by ozone, nitrogen dioxide, and carbon monoxide. The predicted AQI values closely align with observed measurements, demonstrating strong generalizability across regions. An interactive dashboard was developed to visualize AQI predictions and pollutant contributions across countries, improving practical usability for environmental monitoring. Overall, this study provides a comprehensive framework for global AQI prediction and demonstrates the potential of machine learning to support decision-making in environmental management and public health planning.

Day by day the air pollution becomes serious concern in India as well as in overall world. Proper or accurate prediction or forecast of Air Quality or the concentration level of other Ambient air pollutants such as Sulfur Dioxide, Nitrogen Dioxide, Carbon Monoxide, Particulate Matter having diameter less than 10µ, Particulate Matter having diameter less than 2.5µ, Ozone, etc. is very important because impact of these factors on human health becomes severe. This literature review focuses on the various techniques used for prediction or modelling of Air Quality Index (AQI) and forecasting of future concentration levels of pollutants that may cause the air pollution so that governing bodies can take the actions to reduce the pollution.

During the last decades, air pollution has become a serious environmental hazard. Its impact on public health and safety, as well as on the ecosystem, has been dramatic. Forecasting the levels of air pollution to maintain the climatic conditions and environmental protection becomes crucial for government authorities to develop strategies for the prevention of pollution. This study aims to evaluate the atmospheric air pollution of the city of Zahleh located in the geographic zone of Bekaa. The study aims to determine a relationship between variations in ambient particulate concentrations during a short time. The data was collected from June 2017 to June 2018. In order to predict the Air Quality Index (AQI), Naïve, Exponential Smoothing, TBATS (a forecasting method to model time series data), and Seasonal Autoregressive Integrated Moving Average (SARIMA) models were implemented. The performance of these models for predicting air quality is measured using the Mean Absolute Error (MAE), the Root Mean Square Error (RMSE), and the Relative Error (RE). SARIMA model is the most accurate in prediction of AQI (RMSE = 38.04, MAE = 22.52 and RE = 0.16). The results reveal that SARIMA can be applied to cities like Zahleh to assess the level of air pollution and to prevent harmful impacts on health. Furthermore, the authorities responsible for controlling the air quality may use this model to measure the level of air pollution in the nearest future and establish a mechanism to identify the high peaks of air pollution.

Accurate rainfall forecasting is crucial in activities such as agricultural planning, water resource management, and flood preparedness, so failure to accurately forecast rainfall might undermine these activities. This study aimed to improve the accuracy of daily rainfall prediction in Hai district, Kilimanjaro region, Tanzania, a region characterised by a bimodal rainfall pattern, by comparing rainfall prediction accuracies of Traditional Machine Learning (TML) and Deep Learning (DL) models. The 25-year daily rainfall dataset for Hai district was first pre-processed by cleaning, normalisation, and splitting into training (first 60%), validation (next 20%), and test (last 20%) sets. TML models (K-Nearest Neighbours (KNN) through its variant KNN Regressor and Support Vector Machine (SVM) through its variant Support Vector Regressor (SVR)) were developed, trained, and their performances compared with developed and trained DL models (Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU)). Each model’s test performance was evaluated using the Root Mean Squared Error (RMSE) metric. The findings revealed that the DL models, LSTM and GRU, achieved test RMSE scores of 6.35 mm and 6.49 mm, respectively, while the TML models, KNN and SVM, achieved test RMSE scores of 6.61 mm and 10.42 mm, respectively. The study demonstrated that DL models are more effective for accurate rainfall forecasting than TML models in regions with complex rainfall patterns. The insights gained will guide the development of more reliable, location-specific weather prediction systems, ultimately supporting informed decision-making in agriculture, environmental management, and flood risk reduction in Tanzania and other regions facing similar climatic challenges

Prediction of air quality index for certain locations in Nigeria has been carried out. The data used for this research was gotten from purple air data web site. The locations selected are; University of Ibadan, Edo Broadcasting station, Kogi, Nimet, Benin, Lekki phase1 in Lagos and Space center, Kebbi. Air quality Index was calculated based on PM2.5 data obtained from the purple air website. The other variables obtained are PM10, PM1, Temperature and Humidity. The machine learning algorithms used for modelling are linear regression, support vector machines (SVM), K-Nearest Neighbors (KNN), and regression trees. These algorithms were obtained using R programming language. The results show that for University of Ibadan station. AQI increased towards the end of 2021 up till the end of February 2022. There was a steady decrease in AQI towards the end of May 2022. For Edo Broadcasting station, AQI increased towards the end of 2021 up till the end of February 2022. There was a steady decrease in AQI towards the end of May 2022. There was an increase after May 2022. For Kogi station, AQI increased towards the end of October 2021 up till the end of February 2022. There was a steady decrease in AQI in April 2022. For Lekki phase 1, Lagos station, AQI increased from April 2021 towards the end of December 2021. There was a steady decrease in AQI towards the end of April 2022. For Nimet, Benin station, AQI increased steadily from October 2021 up till the end of February 2022. There was a steady decrease in AQI towards the end of April 2022 and a slight increase in May 2022. For Space center, Kebbi station, AQI decreased steadily from August 2021 up till the end of February 2022. There was a steady increase of AQI in March 2022 and a slight decrease in May 2022. AQI shows that the quality of air in Ibadan was classified as good, Edo broadcasting and Nimet, Benin were unhealthy for sensitive groups while Kogi, Space center Kebbi and Lagos were classified as moderate. Modelling result shows that of all the machine learning model selected for this research work, K-nearest neighbors was found to be the best model for the prediction of AQI across Nigeria. This is because it had the smallest value for the errors. Hence for the prediction of AQI in Nigeria, KNN is recommended.

One of the most burning environmental issues of the contemporary urban and industrial areas has turned out to be air pollution which poses serious threats to human health, ecological balance, and overall sustainability. The traditional monitoring stations are very precise, but expensive, sparsely distributed geographically and cannot give high spatial temporal information. To overcome these constraints, the presented study suggests to simulate and design an Internet of Things (IoT)-based air quality monitoring system with the use of low-cost and smart sensors and use MATLAB/Simulink as the main development environment. The suggested framework will incorporate particulate matter (PM2.5 and PM10), nitrogen dioxide (NO2), carbon monoxide (CO), ozone (O3), and volatile organic compound (VOC) sensors. The nodes have a microcontroller unit (ESP32 or Raspberry Pi), wireless communication system (Wi-Fi, LoRa, or NB-IoT), and an energy management system powered by a solar to maintain sustainability. A variety of data pre-processing methods (calibration, temperature and humidity compensation, noise elimination with digital filters, and Kalman filtering) are performed at the edge level to improve measurement accuracy prior to transmission, followed by the simulation of the end-to-end system (data acquisition and transmission, cloud integration and visualization) with the help of MATLAB/Simulink. Moreover, more complex algorithms like sensor fusion and machine learning models (Artificial Neural Networks (ANNs) and Long Short-Term Memory (LSTM) networks) are used to enhance air quality index (AQI) prediction and allow making predictions in the short term on pollution. The system is evaluated by use of performance metrics like: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), correlation coefficient (R2), energy consumption and network latency to determine the anticipated outcome of this research; a scalable, cost-effective, and energy-efficient IoT network that will be able to offer reliable real-time air quality monitoring and forecasting. The solution is a part of the smart city plans, aids policy-making regarding the environment, and contributes to the increase of the public awareness of their health through providing them with the accessible and high-quality air quality data.

Accurate forecasting of diabetes burden is essential for healthcare planning, resource allocation, and policy-making. While deep learning models have demonstrated superior predictive capabilities, their real-world applicability is constrained by computational complexity and data quality challenges. This study evaluates the trade-offs between predictive accuracy, robustness, and computational efficiency in diabetes forecasting. Four forecasting models were selected based on their ability to capture temporal dependencies and handle missing healthcare data: Transformer with Variational Autoencoder (VAE), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and AutoRegressive Integrated Moving Average (ARIMA). Annual data on Disability-Adjusted Life Years (DALYs), Deaths, and Prevalence from 1990 to 2021 were used to train (1990–2014) and evaluate (2015–2021) the models. Performance was measured using Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). Robustness tests introduced noise and missing data, while computational efficiency was assessed in terms of training time, inference speed, and memory usage. Statistical significance was analyzed using ANOVA and Tukey’s post-hoc tests. The Transformer-VAE model achieved the highest predictive accuracy (MAE: 0.425, RMSE: 0.501) and demonstrated superior resilience to noisy and incomplete data (p < 0.01). LSTM effectively captured short-term patterns but struggled with long-term dependencies, while GRU, though computationally efficient, exhibited higher error rates. ARIMA, despite being resource-efficient, showed limited capability in modeling long-term trends, indicating potential benefits in hybrid approaches. While Transformer-VAE provides the most accurate diabetes burden forecasting, its high computational cost and interpretability challenges limit its scalability in resource-constrained settings. These findings highlight the potential of deep learning models for healthcare forecasting, while underscoring the need for further validation before integration into real-world public health decision-making.

Air pollution is a serious issue that affects many people around the globe. It has many negative effects, especially on health, and is measured by the air quality index (AQI). India is one of the most polluted countries in the world, with over 660 million people living in areas with air pollution above the standard AQI. We hypothesize that traditional time series processing algorithms for forecasting, such as the Seasonal Autoregressive Integrated Moving Average (SARIMA), capture seasonal variations and can forecast future AQI levels better than the more common complex deep-learning models, like long short-term memory (LSTM) models. We used a dataset from the Central Pollution Control Board, the official portal of the Government of India that contains time series data for different cities. We created a forecasting model using the SARIMA and LSTM models. Our findings reveal that the SARIMA model effectively captures seasonal patterns in the data for all cities, except Chennai, predicting values with a minimal error margin. In contrast, the LSTM model, while comparable in some cases, generally exhibits poorer performance across more cities and underperforms compared to SARIMA even in its better scenarios. This trend is further evidenced by the root-mean-squared error (RMSE) results, where SARIMA consistently outperforms LSTM in all cities. Overall, our methodology demonstrates high accuracy, holding significant potential to positively impact numerous lives, and supports our hypothesis.

Accurate air quality prediction is crucial for environmental monitoring and public health. This study explores a novel approach using machine learning algorithms and large language models to predict the Air Quality Index (AQI). While traditional AQI prediction relies on complex models and extensive data on pollutants and meteorological factors, this research investigates the use of readily available fuel consumption data as an alternative predictor, given its close link to emissions. The study employs supervised machine learning algorithms, including Random Forest and Gradient Boosting, utilizing fuel consumption data as input features to build predictive models. Additionally, a state-of-the-art large language model, GPT-3.5-turbo-instruct, is fine-tuned on historical AQI data and evaluated for its predictive capabilities. The performance of both machine learning models and the language model is compared using various metrics, and the results demonstrate that both approaches achieve high AQI prediction accuracy, outperforming traditional methods based on pollutant concentration data. Notably, the fine-tuned language model exhibits superior performance, potentially due to its ability to capture complex dependencies and contextual information from the training data. This work highlights the potential of leveraging readily available fuel consumption data and advanced language models for accurate and cost-effective AQI prediction. The findings have significant implications for developing scalable air quality monitoring systems, enabling timely interventions and informed decision-making to mitigate the adverse effects of air pollution.

In smart cities, air pollution has harmful impacts on human physical health and the quality of living environment. correctly predicting air quality is important for developing effective strategies to reduce air pollution and promote healthier, more sustainable environments. Tracking and predicting air pollution is essential for enabling individuals to make wellinformed choices that safeguard their health. Predicting air quality is vital for public health, environmental management, and the development of effective policies. This research focuses on predicting the Air Quality Index (AQI) using machine learning techniques, with an emphasis on improving model efficiency and prediction accuracy. This study compares the performance of two regression algorithms: Linear Regression and Principal Component Regression (PCR) with Decision Tree Regression, across several key evaluation metrics. The performance is assessed using measures such as Mean Squared Error (MSE), Mean Percentage Error (MPE), Mean Absolute Percentage Error (MAPE), Median Absolute Error (MedAE), Explained Variance, Adjusted R². The analysis reveals that the PCR with Decision Tree Regression model outperforms Linear Regression in terms of accuracy, as indicated by lower error values and higher explained variance. The superior model also demonstrates better generalization, with more robust metrics like MedAE, which reduces sensitivity to outliers. Overall, the study highlights the advantages of combining principal component regression with decision tree regression for enhanced predictive accuracy.

Fractional-order models (FOMs) have been recognized as superior tools for capturing the complex electrochemical dynamics of lithium-ion batteries, outperforming integer-order models in accuracy, robustness, and adaptability. Parameter identification (PI) is essential for FOMs, as its accuracy directly affects the model’s ability to predict battery behavior and estimate critical states such as state of charge (SOC) and state of health (SOH). In this study, a hybrid deep learning approach has been introduced for FOM PI, representing the first application of deep learning in this domain. A simulation platform was developed to generate datasets using Sobol and Monte Carlo sampling methods. Five deep learning models were constructed: long short-term memory (LSTM), gated recurrent unit (GRU), one-dimensional convolutional neural network (1DCNN), and hybrid models combining 1DCNN with LSTM and GRU. Hyperparameters were optimized using Optuna, and enhancements such as Huber loss for robustness to outliers, stochastic weight averaging, and exponential moving average for training stability were incorporated. The primary contribution lies in the hybrid architectures, which integrate convolutional feature extraction with recurrent temporal modeling, outperforming standalone models. On a test set of 1000 samples, the improved 1DCNN + GRU model achieved an overall root mean square error (RMSE) of 0.2223 and a mean absolute percentage error (MAPE) of 0.27%, representing nearly a 50% reduction in RMSE compared to its baseline. This performance surpasses that of the improved LSTM model, which yielded a MAPE of 9.50%, as evidenced by tighter scatter plot alignments along the diagonal and reduced error dispersion in box plots. Terminal voltage prediction was validated with an average RMSE of 0.002059 and mean absolute error (MAE) of 0.001387, demonstrating high-fidelity dynamic reconstruction. By advancing data-driven PI, this framework is well-positioned to enable real-time integration into battery management systems.

The impact of air pollution on health is measured by the Air Quality Index (AQI). Accurate AQI prediction is essential for pollution reduction and public health recommendations. Traditional methods of monitoring air quality are inaccurate and time-consuming. This study uses IoT-based air quality data from Kampung Kalipaten, Tangerang to build an AQI prediction model with machine learning, specifically an ensemble model. Ensemble techniques such as bagging and boosting, which increase the reliability of predictions by reducing model bias and inconsistency, improve AQI prediction. Four ensemble models used in this study, they are Random Forest Regressor, Gradient Boosting Regressor, Adaboosting Regressor, and Bagging Regressor. As the evaluation, RMSE and R2 metrics used. Random Forest Regressor perform the best with RMSE value of 0.6054 and R2 value of 0.6271, although no significant differences of RMSE and R2 value of the rest models.

Air-quality monitoring and forecasting are crucial for atmosphere pollution control and management. We propose an innovative data-driven framework for air quality index (AQI) prediction by integrating GNSS radio occultation (GNSS-RO) observation and weather modeling. Empowered by the state-of-the-art machine learning approach, our method can effectively predict regional AQI with a comparable accuracy much more quickly than the traditional numerical modeling and simulation approach. In a real case study using a representative region of China, our data-driven approach achieves a 2000 times speedup; meanwhile, the prediction error measured by rRMSE is only 2.4%. We investigate further the effects of different models, hyperparameters, and meteorological factors on the performance of our AQI prediction framework, and reveal that wind field and atmospheric boundary-layer height are important influencing factors of AQI. This paper showcases a direct application of GNSS-RO observation in assisting in forecasting regional AQI. From a machine learning point of view, it provides a new way to leverage the unique merits of GNSS atmospheric remote sensing technology with the help of the more traditional weather forecasting modeling approach.

Potatoes are an important crop in the world; they are the main source of food for a large number of people globally and also provide an income for many people. The true forecasting of potato yields is a determining factor for the rational use and maximization of agricultural practices, responsible management of the resources, and wider regions’ food security. The latest discoveries in machine learning and deep learning provide new directions to yield prediction models more accurately and sparingly. From the study, we evaluated different types of predictive models, including K-nearest neighbors (KNN), gradient boosting, XGBoost, and multilayer perceptron that use machine learning, as well as graph neural networks (GNNs), gated recurrent units (GRUs), and long short-term memory networks (LSTM), which are popular in deep learning models. These models are evaluated on the basis of some performance measures like mean squared error (MSE), root mean squared error (RMSE), and mean absolute error (MAE) to know how much they accurately predict the potato yields. The terminal results show that although gradient boosting and XGBoost algorithms are good at potato yield prediction, GNNs and LSTMs not only have the advantage of high accuracy but also capture the complex spatial and temporal patterns in the data. Gradient boosting resulted in an MSE of 0.03438 and an R2 of 0.49168, while XGBoost had an MSE of 0.03583 and an R2 of 0.35106. Out of all deep learning models, GNNs displayed an MSE of 0.02363 and an R2 of 0.51719, excelling in the overall performance. LSTMs and GRUs were reported to be very promising as well, with LSTMs comprehending an MSE of 0.03177 and GRUs grabbing an MSE of 0.03150. These findings underscore the potential of advanced predictive models to support sustainable agricultural practices and informed decision-making in the context of potato farming.