Search Results (2,954)

Air pollution, particularly particulate matter (PM₁, PM_2.5, and PM₁₀), poses a significant environmental health risk globally. This study evaluates the predictive performance of three machine learning algorithms, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Random Forest (RF), for forecasting particulate matter concentrations in four Brazilian cities (Porto Alegre, Recife, Goiânia, and Belém), which share similar demographic and urbanization characteristics but differ in geographic and climatic conditions. Using data from the Copernicus Atmosphere Monitoring Service, daily concentrations of PM₁, PM_2.5, and PM₁₀ were modeled based on meteorological variables, including air temperature, relative humidity, wind speed, atmospheric pressure, and accumulated precipitation. The models were tested under two climate change scenarios (+2 °C and +4 °C temperature increases). The results indicate that RF consistently outperformed the other models, achieving low RMSE values, around 0.3 µg/m³, across all cities, regardless of their geographic and climatic differences. KNN showed stable performance under moderate temperature increases (+2 °C) but exhibited higher errors under more extreme warming, while SVM demonstrated higher sensitivity to temperature changes, leading to greater variability in bivariate contexts. However, in multivariate contexts, SVM adjusted better, improving its predictive performance by accounting for the combined influence of multiple meteorological variables. These findings underscore the importance of selecting suitable machine learning models, with RF proving to be the most robust approach for particulate matter prediction across diverse environmental contexts. This study contributes valuable insights for the development of region-specific air quality management strategies in the face of climate change. Full article

This study investigates the financial and operational feasibility of deploying electric vehicle (EV) charging infrastructure within high-density residential buildings, utilizing empirical operational data combined with comprehensive financial modeling. A 14-day monitoring period conducted at a residential complex comprising 958 units revealed distinct charging behaviors, with demand peaking during weekday evenings between 19:00 and 22:00 and displaying more dispersed yet lower overall utilization during weekends. Energy efficiency emerged as a significant operational constraint, as standby power consumption contributed substantially to total energy losses. Specifically, while total energy consumption reached 248.342 kW, only 138.24 kW were directly delivered to users, underscoring the necessity for energy-efficient hardware and intelligent load management systems to minimize idle consumption. The financial analysis identified pricing as the most critical determinant of project viability. Under current cost structures, financial break-even was attainable only at a profit margin of 0.2286 USD (8 THB) per kWh, while lower margins resulted in persistent financial deficits. Sensitivity analysis further demonstrated the considerable vulnerability of the project’s financial performance to small fluctuations in profit share and utilization rate. A 10% reduction in either parameter entirely eliminated the project’s ability to reach payback, while variations in energy costs, capital expenditures (CAPEX), and operational expenditures (OPEX) exerted comparatively limited influence. These findings emphasize the importance of precise demand forecasting, adaptive pricing strategies, and proactive government intervention to mitigate financial risks associated with residential EV charging deployment. Policy measures such as capital subsidies, technical regulations, and transparent pricing frameworks are essential to incentivize private sector investment and support sustainable expansion of EV infrastructure in residential sectors. Full article

Coal is a vital part of China’s energy system, and accurately predicting mine water inflow is crucial for ensuring the safety and efficiency of coal mining. To enhance prediction accuracy, this study introduces a hybrid model—CEEMDAN-OVMD-Transformer—that combines Adaptive Noise Complete Ensemble Empirical Mode Decomposition (CEEMDAN), Optimal Variational Mode Decomposition (OVMD), and the Transformer architecture. This combined model is used to forecast water inflow at the Heidaigou Coal Mine. The experimental results show that the proposed model achieves excellent predictive accuracy, with a Mean Absolute Error (MAE) of 0.507, a Root Mean Square Error (RMSE) of 0.614, a Mean Absolute Percentage Error (MAPE) of 0.010, and a Coefficient of Determination (R²) of 0.948. Compared to the standalone Transformer model, the CEEMDAN-OVMD-Transformer model reduces the MAE by 34.50% and increases the R² by approximately 3.04%, indicating a significant improvement in forecasting accuracy. The findings demonstrate that the CEEMDAN-OVMD-Transformer hybrid model can effectively reduce the complexity of high-frequency components in mine water inflow time series, thereby enhancing the stability and reliability of predictions. This research presents a new and effective approach for mine water inflow forecasting and offers valuable technical guidance for water hazard prevention and control in similar coal mining environments. Full article

Accurate modeling of car-following behavior is essential for understanding traffic dynamics and enabling predictive control in intelligent transportation systems. This study presents a novel data-driven framework that combines information-theoretic input selection via conditional transfer entropy (CTE) with dynamic mode decomposition with control (DMDc) for identifying and forecasting car-following dynamics. In the first step, CTE is employed to identify the specific vehicles that exert directional influence on a given subject vehicle, thereby systematically determining the relevant control inputs for modeling its behavior. In the second step, DMDc is applied to estimate and predict the dynamics by reconstructing the closed-form expression of the dynamical system governing the subject vehicle’s motion. Unlike conventional machine learning models that typically seek a single generalized representation across all drivers, our framework develops individualized models that explicitly preserve driver heterogeneity. Using both synthetic data from multiple traffic models and real-world naturalistic driving datasets, we demonstrate that DMDc accurately captures nonlinear vehicle interactions and achieves high-fidelity short-term predictions. Analysis of the estimated system matrices reveals that DMDc naturally approximates kinematic relationships, further reinforcing its interpretability. Importantly, this is the first study to apply DMDc to model and predict car-following behavior using real-world driving data. The proposed framework offers a computationally efficient and interpretable tool for traffic behavior analysis, with potential applications in adaptive traffic control, autonomous vehicle planning, and human-driver modeling. Full article

This review examines South Africa’s escalating flood vulnerability through a synthesis of over 80 peer-reviewed articles, historical records, policy reports, and case studies. Using a PRISMA-guided analysis, the study identifies key climatic drivers, including extreme rainfall from tropical–temperate interactions, cut-off lows, and La Niña conditions that interact with structural weaknesses such as inadequate drainage, poorly maintained stormwater systems, and rapid urban expansion. Apartheid-era spatial planning has further entrenched risk by locating marginalised communities in floodplains. Governance failures like weak disaster risk reduction (DRR) policies, fragmented institutional coordination, and insufficient early warning systems intensify flood vulnerabilities. Catastrophic events in KwaZulu-Natal (KZN) and the Western Cape (WC) illustrate the consequences exemplified by the April 2022 KZN floods alone, which caused over 450 deaths, displaced more than 40,000 people, and generated damages exceeding ZAR 17 billion. Nationally, more than 1500 flood-related fatalities have been documented in the past two decades. Emerging resilience pathways include ecosystem-based adaptation, green infrastructure, participatory governance, integration of Indigenous knowledge, improved hydrological forecasting, and stricter land-use enforcement. These approaches can simultaneously reduce physical risks and address entrenched socio-economic inequalities. However, significant gaps remain in spatial flood modelling, gender-sensitive responses, urban–rural disparities, and policy implementation. The review concludes that South Africa urgently requires integrated, multi-scalar strategies that combine scientific innovation, policy reform, and community-based action. Embedding these insights into disaster management policy and planning is essential to curb escalating losses and build long-term resilience in the face of climate change. Full article

Modern artificial intelligence methods are increasingly applied in hydrology, particularly for forecasting water inflow into reservoirs. However, their limited interpretability constrains practical deployment in critical water resource management systems. Explainable AI offers solutions aimed at increasing the transparency of models, which makes the topic relevant in the context of developing sustainable and trusted AI systems in hydrology. Articles published in leading scientific journals in recent years were selected for the review. The selection criteria were the application of XAI methods in hydrological forecasting problems and the presence of a quantitative assessment of interpretability. The main attention is paid to approaches combining LSTM, GRU, CNN, and ensembles with XAI methods such as SHAP, LIME, Grad-CAM, and ICE. The results of the review show that XAI mechanisms increase confidence in AI forecasts, identify important meteorological features, and allow analyzing parameter interactions. However, there is a lack of standardization of interpretation, especially in problems with high-dimensional input data. The review emphasizes the need to develop robust, unified XAI approaches that can be integrated into next-generation hydrological models. Full article

This study aims to guide the management and service of highways towards a more efficient and intelligent direction, and also provides intelligent and green data support for achieving sustainable development goals. The forecasting of traffic flow at highway stations serves as the cornerstone for spatiotemporal analysis and is vital for effective highway management and control. Despite considerable advancements in data-driven traffic flow prediction, the majority of existing models fail to differentiate between directions. Specifically, entrance flow prediction has applications in dynamic route guidance, disseminating real-time traffic conditions, and offering optimal entrance selection suggestions. Meanwhile, exit flow prediction is instrumental for congestion and accident alerts, as well as for road network optimization decisions. In light of these needs, this study introduces an enhanced heterogeneous spatiotemporal graph network model tailored for predicting highway station traffic flow. To accurately capture the dynamic impact of upstream toll stations on the target station’s flow, we devise an influence probability matrix. This matrix, in conjunction with the covariance matrix across toll stations, updated graph structure data, and integrated external weather conditions, allows the attention mechanism to assign varied combination weights to the target toll station from temporal, spatial, and external standpoints, thereby augmenting prediction accuracy. We undertook a case study utilizing traffic flow data from the Chengdu-Chengyu station on the Sichuan Highway to gauge the efficacy of our proposed model. The experimental outcomes indicate that our model surpasses other baseline models in performance metrics. This study provides valuable insights for highway management and control, as well as for reducing traffic congestion. Furthermore, this research highlights the importance of using data-driven approaches to reduce carbon emissions associated with transportation, enhance resource allocation at toll plazas, and promote sustainable highway transportation systems. Full article

Crop aeroponic cultivation still faces issues such as insufficient precision in water supply control and scientifically-based irrigation scheduling. To address this challenge, the present study aims to establish a precision irrigation protocol adapted to the characteristics of crop aeroponic cultivation. Using coriander (Coriandrum sativum L.) as the experimental subject, crop water requirements were estimated utilizing both the FAO56 P-M equation and its revised form. The RMSE between the water requirement measured values and the calculated values using the P-M formula is 2.12 mm, the MAE is 2.0 mm, and the MAPE is 14.29%. The RMSE between the water requirement measured values and the calculated values using the revised P-M formula is 0.88 mm, the MAE is 0.82 mm, and the MAPE is 5.78%. The results indicate that the water requirement values calculated using the revised P-M formula are closer to the measured values. For model development, this study used coriander evapotranspiration as a basis. Major environmental variables influencing water requirement were selected as input features, and the daily reference water requirement served as the output. Three modeling approaches were implemented: Random Forest (RF), Bagging, and M5P Model Tree algorithms. The results indicate that, in comparing various input combinations (C1: air temperature, relative humidity, atmospheric pressure, wind speed, radiation, photoperiod; C2: air temperature, relative humidity, wind speed, radiation; C3: air temperature, relative humidity, radiation), the RF model based on C1 input demonstrated superior performance with RMSE = 0.121 mm/d, MAE = 0.134 mm/d, MAPE = 2.123%, and R² = 0.971. It significantly outperforms the RF models with other input combinations, as well as the Bagging and M5P models across all input scenarios, in terms of convergence rate, determination coefficient, and comprehensive performance. Its predictions aligned more closely with observed data, showing enhanced accuracy and adaptability. This optimized prediction model demonstrates particular suitability for forecasting water requirements in aeroponic coriander production and provides theoretical support for efficient, intelligent water-saving management in crop aeroponic cultivation. Full article

Time series clustering plays a vital role in various analytical and pattern recognition tasks by partitioning structurally similar sequences into semantically coherent groups, thereby facilitating downstream analysis. However, building high-quality clustering models remains challenging due to three key issues: (i) capturing dynamic shape variations across sequences, (ii) ensuring discriminative cluster structures, and (iii) enabling end-to-end optimization. To address these challenges, we propose KDiscShapeNet, a structure-aware clustering framework that systematically extends the classical k-Shape model. First, to enhance temporal structure modeling, we adopt Kolmogorov–Arnold Networks (KAN) as the encoder, which leverages high-order functional representations to effectively capture elastic distortions and multi-scale shape features of time series. Second, to improve intra-cluster compactness and inter-cluster separability, we incorporate a dual-loss constraint by combining Center Loss and Supervised Contrastive Loss, thus enhancing the discriminative structure of the embedding space. Third, to overcome the non-differentiability of traditional K-Shape clustering, we introduce Differentiable k-Shape, embedding the normalized cross-correlation (NCC) metric into a differentiable framework that enables joint training of the encoder and the clustering module. We evaluate KDiscShapeNet on nine benchmark datasets from the UCR Archive and the ETT suite, spanning healthcare, industrial monitoring, energy forecasting, and astronomy. On the Trace dataset, it achieves an ARI of 0.916, NMI of 0.927, and Silhouette score of 0.931; on the large-scale ETTh1 dataset, it improves ARI by 5.8% and NMI by 17.4% over the best baseline. Statistical tests confirm the significance of these improvements (p < 0.01). Overall, the results highlight the robustness and practical utility of KDiscShapeNet, offering a novel and interpretable framework for time series clustering. Full article

Wind turbine power forecasting is crucial for optimising energy production, planning maintenance, and enhancing grid stability. This research focuses on predicting the output of a Senvion MM92 wind turbine at the Kelmarsh wind farm in the UK using SCADA data from 2020. Two approaches are explored: a hybrid model combining Stochastic Differential Equations (SDEs) with Neural Networks (NNs) and Deep Learning models, in particular, Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), and the Combination of Convolutional Neural Networks (CNNs) and LSTM. Notably, while SDE-NN models are well suited for predictions in cases where data patterns are chaotic and lack consistent trends, incorporating stochastic processes increases the complexity of learning within SDE models. Moreover, it is worth mentioning that while SDE-NNs cannot be classified as purely “white box” models, they are also not entirely “black box” like traditional Neural Networks. Instead, they occupy a middle ground, offering improved interpretability over pure NNs while still leveraging the power of Deep Learning. This balance is precious in fields such as wind power prediction, where accuracy and understanding of the underlying physical processes are essential. The evaluation of the results demonstrates the effectiveness of the SDE-NNs compared to traditional Deep Learning models for wind power prediction. The SDE-NNs achieve slightly better accuracy than other Deep Learning models, highlighting their potential as a powerful alternative. Full article

This study addresses the critical need for accurate tourism demand (TD) forecasting in Bulgaria using economic indicators, developing robust predictive models to navigate post-pandemic market volatility. The COVID-19 pandemic exposed tourism’s vulnerability to systemic shocks, highlighting deficiencies in traditional forecasting approaches. Bulgaria’s tourism industry, characterized by strong seasonal variations and economic sensitivity, requires enhanced methodologies for strategic planning in uncertain environments. The research employs comprehensive comparative analysis of machine learning (ML) and deep machine learning (DML) methodologies. Monthly overnight stay data from Bulgaria’s National Statistical Institute (2005–2024) were integrated with COVID-19 case data, Consumer Price Index (CPI) and Bulgarian Gross Domestic Product (GDP) variables for the same period. Multiple approaches were implemented including Prophet with external regressors, Ridge regression, LightGBM, and gradient boosting models using inverse MAE weighting optimization, alongside deep learning architectures such as Bidirectional LSTM with attention mechanisms and XGBoost configurations, as each model statistical significance was estimated. Contrary to prevailing assumptions about deep learning superiority, traditional machine learning ensemble approaches demonstrated superior performance. The ensemble model combining Prophet, LightGBM, and Ridge regression achieved optimal results with MAE of 156,847 and MAPE of 14.23%, outperforming individual models by 10.2%. Deep learning alternatives, particularly Bi-LSTM architectures, exhibited significant deficiencies with negative R² scores, indicating fundamental limitations in capturing seasonal tourism patterns, probable data dependence and overfitting. The findings, provide tourism stakeholders and policymakers with empirically validated forecasting tools for enhanced decision-making. The ensemble approach combined with statistical significance testing offers improved accuracy for investment planning, marketing budget allocation, and operational capacity management during economic volatility. Economic indicator integration enables proactive responses to market disruptions, supporting resilient tourism planning strategies and crisis management protocols. Full article

The karst terrain of Guangxi, China, characterized by steep slopes and thin residual soils, is highly vulnerable to rainfall-induced shallow landslides. Timely and accurate displacement forecasting is critical for early warning and risk mitigation. However, most existing systems depend on centralized computation, leading to latency and reduced responsiveness. Moreover, conventional forecasting models are often too computationally intensive for edge devices with limited processing resources. To address these constraints, we present EoML-SlideNet, a lightweight forecasting framework designed for resource-limited hardware. It decomposes displacement and triggers into trend and periodic components, then applies the Dual-Band Lasso-Enhanced Latent Variable (DBLE–LV) module to select compact, interpretable features via cross-correlation, LASSO, and VIF screening. A small autoregressive model predicts the trend, while a lightweight neural network captures periodic fluctuations. Their outputs are combined to estimate displacement. All models were evaluated on a single CPU-only workstation to ensure fair comparison. This study introduces floating-point operations (FLOPs), alongside runtime, as practical evaluation metrics for landslide displacement prediction models. A site-specific multi-sensor dataset was developed to monitor rainfall-triggered landslide behavior in the karst terrain of Guangxi. The experimental results show that EoML-SlideNet achieves 2–4 times lower MAE/RMSE than the most accurate deep learning and the lightest baseline models, while offering 3–30 times faster inference. These results demonstrate that low-complexity models can match or surpass the accuracy of deep networks while achieving latency and FLOP levels suitable for edge deployment without dependence on remote servers. Full article

High energy-consuming industrial parks are characterized by high base-load-to-peak-valley ratios, overlapping production cycles, and megawatt-scale step changes, which significantly complicate short-term load forecasting. To tackle these challenges, this study proposes a novel forecasting framework that combines hierarchical multimodal decomposition with a hybrid deep learning architecture. First, Maximal Information Coefficient (MIC) analysis is applied to identify key input features and eliminate redundancy. The load series is then decomposed in two stages: seasonal-trend decomposition uses the Loess (STL) isolates trend and seasonal components, while variational mode decomposition (VMD) further disaggregates the residual into multi-scale modes. This hierarchical approach enhances signal clarity and preserves temporal structure. A parallel neural architecture is subsequently developed, integrating an Informer network to model long-term trends and a bidirectional gated recurrent unit (BiGRU) to capture short-term fluctuations. Case studies based on real-world load data from a typical industrial park in northeastern China demonstrate that the proposed model achieves significantly improved forecasting accuracy and robustness compared to benchmark methods. These results provide strong technical support for fine-grained load prediction and intelligent dispatch in high energy-consuming industrial scenarios. Full article

This study proposes a hybrid forecasting framework that integrates the Auto-Regressive Integrated Moving Average (ARIMA) model with multiple variations of the Ata model, using a performance-based model selection strategy to enhance forecasting accuracy on the M3 and M4 competition datasets. For each time series, seven versions of the Ata model are generated by adjusting level and trend parameters, and the version with the lowest in-sample symmetric mean absolute percentage error (sMAPE) is selected. To improve robustness and prevent overfitting, the median-performing Ata model is also included. These selected models’ forecasts are then combined with ARIMA outputs through optimized weighting schemes tailored to the characteristics of each series. Given the varying frequencies (e.g., yearly, quarterly, monthly, weekly, daily, and hourly) and diverse lengths of time series, a grid search algorithm is employed to determine the best hybrid combination for each frequency group. The model is applied in a series-specific manner, allowing it to adapt to different seasonal, trend, and irregular patterns. Extensive empirical results demonstrate that the hybrid model outperforms its individual components and traditional benchmarks across all frequency categories. It ranked first in the M3 competition and achieved second place in the M4 competition based on the official error metric, the sMAPE and Overall Weighted Average (OWA), respectively. The results highlight the framework’s adaptability and scalability for complex, heterogeneous time series environments. Full article

Despite significant progress, precipitation forecasting in West Africa remains challenging due to the complexity of atmospheric processes and the region’s climatic variability. This study aims to identify optimal configurations of the WRF model to improve precipitation forecasting. To evaluate the sensitivity of the model’s physical parameterizations, 15 configurations were tested by combining various cumulus parameterization schemes (CPSs) and planetary boundary layer (PBL) schemes. The analysis examines two contrasting rainfall events in Senegal: one characterized by widespread intense precipitation and another featuring localized moderate rainfall. Simulated rainfall, temperature, and humidity were validated against rain gauges, satellite products (ENACTS, ARC, CHIRPS, and IMERG), and ERA5 reanalysis data. The results show that the WRF configurations achieve correlation coefficients (r) ranging from 0.27 to 0.62 against ENACTS and from 0.15 to 0.41 against rain gauges. The sensitivity analysis reveals that PBL schemes primarily influence temperature and humidity, while CPSs significantly affect precipitation. For the heavy rainfall event, several configurations accurately captured the observed patterns, particularly those using Tiedtke or Grell–Devenyi CPSs coupled with the Mellor–Yamada–Janjic (MYJ) PBL. However, the model showed limited skill in simulating localized convection during the moderate rainfall event. These findings highlight the importance of selecting appropriate parameterizations to enhance WRF-based precipitation forecasting, especially for extreme weather events in West Africa. Full article