Comparative Analysis of Multiple Deep Learning Models for Forecasting Monthly Ambient PM2.5 Concentrations: A Case Study in Dezhou City, China

Atmosphere, Год журнала: 2024, Номер 15(12), С. 1432 - 1432

Опубликована: Ноя. 28, 2024

Ambient air pollution affects human health, vegetative growth and sustainable socio-economic development. Therefore, data in Dezhou City China are collected from January 2014 to December 2023, multiple deep learning models used forecast PM2.5 concentrations. The ability of the is evaluated compared with observed using various statistical parameters. Although all eight can accomplish forecasting assignments, precision accuracy CNN-GRU-LSTM method 34.28% higher than that ANN method. result shows has best performance other seven models, achieving an R (correlation coefficient) 0.9686 RMSE (root mean square error) 4.6491 μg/m3. values CNN, GRU LSTM 57.00%, 35.98% 32.78% method, respectively. results reveal predictor remarkably improves performances benchmark overall forecasting. This research provides a new perspective for predictive ambient model provide scientific basis prevention control.

Язык: Английский

---

An examination of daily CO2 emissions prediction through a comparative analysis of machine learning, deep learning, and statistical models

Environmental Science and Pollution Research, Год журнала: 2025, Номер unknown

Опубликована: Янв. 13, 2025

Abstract Human-induced global warming, primarily attributed to the rise in atmospheric CO 2 , poses a substantial risk survival of humanity. While most research focuses on predicting annual emissions, which are crucial for setting long-term emission mitigation targets, precise prediction daily emissions is equally vital short-term targets. This study examines performance 14 models data from 1/1/2022 30/9/2023 across top four polluting regions (China, India, USA, and EU27&UK). The used include statistical (ARMA, ARIMA, SARMA, SARIMA), three machine learning (support vector (SVM), random forest (RF), gradient boosting (GB)), seven deep (artificial neural network (ANN), recurrent variations such as gated unit (GRU), long memory (LSTM), bidirectional-LSTM (BILSTM), hybrid combinations CNN-RNN). Performance evaluation employs metrics ( R MAE, RMSE, MAPE). results show that (ML) (DL) models, with higher (0.714–0.932) lower RMSE (0.480–0.247) values, respectively, outperformed model, had (− 0.060–0.719) (1.695–0.537) all regions. ML DL was further enhanced by differencing, technique improves accuracy ensuring stationarity creating additional features patterns model can learn. Additionally, applying ensemble techniques bagging voting improved approximately 9.6%, whereas CNN-RNN RNN models. In summary, both relatively similar. However, due high computational requirements associated recommended using bagging. assist accurately forecasting aiding authorities targets reduction.

Язык: Английский

---

The effects of socioeconomic factors on particulate matter concentration in China's: New evidence from spatial econometric model

Journal of Cleaner Production, Год журнала: 2023, Номер 417, С. 137969 - 137969

Опубликована: Июль 3, 2023

Язык: Английский

---

A multi-factor combination prediction model of carbon emissions based on improved CEEMDAN

Environmental Science and Pollution Research, Год журнала: 2024, Номер 31(14), С. 20898 - 20924

Опубликована: Фев. 21, 2024

Язык: Английский

---

An extensive investigation on leveraging machine learning techniques for high-precision predictive modeling of CO ₂ emission

Energy Sources Part A Recovery Utilization and Environmental Effects, Год журнала: 2023, Номер 45(3), С. 9149 - 9177

Опубликована: Июль 9, 2023

Predictive analytics utilizing machine learning algorithms play a pivotal role in various domains, including the profiling of carbon dioxide (CO2) emissions. This research paper delves into an extensive exploration different algorithms, encompassing neural networks with diverse architectures, optimization, training, ensemble, and specialized algorithms. The primary objective this is to evaluate efficacy supervised unsupervised Deep Belief Networks, Feed Forward Neural Gradient Boosting, Regression, as well Convolutional Gaussian, Grey, Markov models, clustering optimization study places particular emphasis on data-driven methodologies cross-validation techniques evaluation models entailing comprehensive validation, testing, employing metrics such R2, MAE, RMSE. employs correlation analysis examine relationship between input parameters emission characteristics. highlights advantageous attributes these accurately forecasting CO2 emissions, evaluating energy sources, improving prediction accuracy, estimating Notably, deep learning, Artificial Networks (ANN), Support Vector Machines (SVM) demonstrate effectiveness across industries, while Modified Regularized Fast Orthogonal-Extreme Learning Machine (MRFO-ELM) algorithm optimizes predictions specifically related coal chemical Hybrid accuracy predicting emissions consumption, whereas gray provide reliable estimates even limited data. However, it important acknowledge certain limitations, data requirements, potential inaccuracies arising from complex factors, constraints faced by developing countries, impact electric vehicle expansion power grid. To optimize survey conducted, involving customization rates, exploring performance model accuracy. outcomes contribute effective monitoring operational environments, thereby aiding executive decision-making processes.

Язык: Английский

---

Modeling and forecasting carbon dioxide emission in Pakistan using a hybrid combination of regression and time series models

Heliyon, Год журнала: 2024, Номер 10(13), С. e33148 - e33148

Опубликована: Июнь 20, 2024

Carbon dioxide (CO

Язык: Английский

---

Controlling carbon emissions through modeling and optimization: addressing an earth system and environment challenge

Modeling Earth Systems and Environment, Год журнала: 2024, Номер 10(5), С. 6003 - 6011

Опубликована: Июль 22, 2024

Abstract This study aims to analyze the trend of carbon dioxide CO 2 emissions from various sources in Pakistan between 1990 and 2020 effectively model underlying dynamics emissions. The design fitting historical data reveal significant trends patterns, highlighting alarming increase These findings underscore necessity for robust policy interventions mitigate achieve sustainable development goals (SDGs). work can contribute addressing challenges recent plans targeting global warming climate emergency. By controlling these parameters, mean reversion be managed, allowing control increasing rate regions threatened by change. O-U provides a valuable framework understanding stochastic nature emissions, offering insights into persistence variability emission levels over time. optimized parametric thresholds model, after synchronizing it with real data, that challenge cannot naturally resolved serious are highly desired. include measures improve air quality, combat

Язык: Английский

---

Prediction of Greenhouse Indoor Air Temperature Using Artificial Intelligence (AI) Combined with Sensitivity Analysis

Horticulturae, Год журнала: 2023, Номер 9(8), С. 853 - 853

Опубликована: Июль 26, 2023

Greenhouses are essential for agricultural production in unfavorable climates. Accurate temperature predictions critical controlling Heating, Ventilation, Air-Conditioning, and Dehumidification (HVACD) lighting systems to optimize plant growth reduce financial losses. In this study, several machine models were employed predict indoor air an even-span Mediterranean greenhouse. Radial Basis Function (RBF), Support Vector Machine (SVM), Gaussian Process Regression (GPR) applied using external parameters such as outside air, relative humidity, wind speed, solar radiation. The results showed that RBF model with the LM learning algorithm outperformed SVM GPR models. had high accuracy reliability RMSE of 0.82 °C, MAPE 1.21%, TSSE 474.07 EF 1.00. prediction can help farmers manage their crops resources efficiently energy inefficiencies lower yields. integration into greenhouse control lead significant savings cost reductions.

Язык: Английский

---

Advanced hybrid neural network techniques for minimizing gas turbine emissions

World Journal of Engineering, Год журнала: 2024, Номер unknown

Опубликована: Окт. 23, 2024

Purpose Emissions have significant environmental impacts. Hence, minimizing emissions is essential. This study aims to use a hybrid neural network model predict carbon monoxide (CO) and nitrogen oxide (NOx) from gas turbines (GTs) enhance emission prediction for GTs in predictive monitoring systems (PEMS). Design/methodology/approach The architecture combines convolutional networks (CNN) bidirectional long-short-term memory (Bi-LSTM) called CNN-BiLSTM with modified extrinsic attention regression. Over five years, data GT power plant was uploaded Google Colab, split into training testing sets (80:20), evaluated using test matrices. model’s performance benchmarked against state-of-the-art methodologies. Findings showed promising results CO NOx emissions. predictions had slight underestimation bias of −0.01, root mean-squared error (RMSE) 0.064, mean absolute (MAE) 0.04 R 2 0.82. an RMSE 0.051, MAE 0.036, 0.887 overestimation +0.01. Research limitations/implications While the demonstrates relative accuracy predictions, there potential further improvement future research. Practical implications Implementing real-time PEMS establishing continuous feedback loop will ensure real-world applications, functioning reduce emissions, fuel consumption running costs. Social Accurate support stricter standards, promote sustainable development goals healthier societal environment. Originality/value paper presents novel approach that integrates CNN Bi-LSTM networks. It considers both spatial temporal mitigate previous shortcomings.

Язык: Английский

---

Carbon Dioxide Emission Forecast: A Review of Existing Models and Future Challenges

Sustainability, Год журнала: 2025, Номер 17(4), С. 1471 - 1471

Опубликована: Фев. 11, 2025

In the face of global climate change, accurately predicting carbon dioxide emissions has become an urgent requirement for environmental science and policy-making. This article provides a systematic review literature on emission forecasting, categorizing existing research into four key aspects. Firstly, regarding model input variables, thorough discussion is conducted pros cons univariate models versus multivariable models, balancing operational simplicity with high accuracy. Secondly, concerning types, detailed comparison made between statistical methods machine learning methods, particular emphasis outstanding performance deep in capturing complex relationships emissions. Thirdly, data, explores annual daily emissions, highlighting practicality predictions policy-making importance providing real-time support policies. Finally, quantity, differences single ensemble are examined, emphasizing potential advantages considering multiple selection. Based literature, future will focus integration multiscale optimizing application in-depth analysis factors influencing prediction, scientific more comprehensive, real-time, adaptive response to challenges change. comprehensive outlook aims provide scientists policymakers reliable information promoting achievement protection sustainable development goals.

Язык: Английский

---