Search Results (603)

Carbon monoxide (CO) remains an important urban air quality and traffic-exposure tracer despite rarely exceeding regulatory limits in modern European cities. This study presents an interpretable machine learning framework for short-term CO forecasting at two operational horizons—next hour and next-day maximum—applied identically to four automated monitoring stations spanning central-urban, urban, urban-background, and suburban typologies in the Greater Athens Area (2021–2024). Using only univariate CO history and calendar-derived features, four learners (support vector regression, random forest, gradient boosting, and a multilayer perceptron) were benchmarked against a naïve baseline under a chronological train/validation/test split. At the next-hour horizon, all learners outperformed naïve at every station, with gradient boosting achieving the best or joint-best skill (test $R^2=0.81$–$0.90$); Wilcoxon signed-rank tests confirmed that these small but consistent margins were statistically significant. The next-day-maximum task proved substantially harder ($R^2=0.44$–$0.59$), with the neural and random forest models overtaking gradient boosting and SVR losing competitiveness. A SHAP (Shapley Additive Explanations) analysis of the next-hour gradient-boosting model showed that the most recent hourly lag dominates the forecast, with an effect nearly an order of magnitude over any other predictor, with hour-of-day encoding and short-lag rolling statistics contributing secondary, sign-consistent effects—providing a transparent, mechanistic account of model behavior rather than a black-box skill score. Unlike ozone, a secondary pollutant whose predictability degrades toward the trafficked urban core, CO concentration and forecastability increase together at the traffic-dominated site, indicating that primary-pollutant forecasts are most reliable precisely where exposure is greatest. These findings support a horizon-specific, interpretable forecasting strategy for operational deployment on real-time monitoring networks. Full article

Predicting electrical demand in distribution systems is a fundamental problem for the efficient operation of smart grids, especially under scenarios of high temporal variability. This study compares two multi-step forecasting strategies for 24-h horizons: a direct 24→24 strategy and a recursive strategy based on sequential 24→1 predictions. Four machine learning and deep learning architectures are evaluated: LSTM, N-HiTS, U-Net, and LightGBM, using real data from electrical feeders belonging to distribution systems in the equatorial region of Ecuador. The methodology includes constructing time windows, non-overlapping train/validation/test partitioning for evaluation, consistent normalization, and comparative analysis using MAE, RMSE, and MAPE metrics. The results show that the direct 24→24 strategy achieves the best overall performance, with LSTM standing out with an approximate MAPE of 4.12%. However, the recursive strategies exhibit greater stability in the face of atypical patterns observed during holidays and weekends. Furthermore, U-Net demonstrates competitive performance in both accuracy and temporal robustness, while LightGBM stands out for its computational efficiency. It is concluded that the selection of a forecasting strategy depends on the required balance between overall accuracy, temporal stability, and computational cost. Full article

Reliable interval forecasting is essential for risk-aware wind power scheduling, yet the strong nonstationarity and complex spatio-temporal coupling of wind power sequences make probabilistic prediction difficult. This study proposes VSTCformer (Variational-mode-decomposition-enhanced Spatio-Temporal Cross-attention transformer), a wind power interval forecasting framework that integrates adaptive variational mode decomposition (VMD) with spatio-temporal cross-attention and multi-quantile regression. The adaptive VMD module automatically determines the number of modes from the uniformity of center-frequency spacing and decomposes the raw power signal into frequency-aligned intrinsic components, while the undecomposed multivariate sequence is preserved to provide global temporal context. Spatial-guided and temporal-modulated attention realize implicit spatio-temporal coupling during encoding, and explicit fusion is achieved through inter-component interaction before the features are passed to the prediction head. Nine quantiles are jointly estimated to produce the median forecast and prediction intervals at multiple confidence levels. Two public benchmarks with markedly different temporal resolution and volatility are used for evaluation. The results show that VSTCformer preserves competitive point-forecast accuracy while substantially improving interval quality: it produces the narrowest prediction intervals at all evaluated horizons and achieves the best coverage-width trade-off in most settings. The central contribution of the framework is therefore an enhanced and well-calibrated uncertainty quantification capability, confirming that decomposition-enhanced spatio-temporal modeling is an effective route for wind power interval forecasting. Full article

Accurate short-term precipitation nowcasting remains challenging in complex terrain regions, where storm displacement, evolution, and orographic enhancement strongly affect precipitation distribution. This study evaluates three precipitation nowcasting schemes initialized from LAPS analysis fields: first-order advection (Control), advection–diffusion (AD), and advection–diffusion coupled with the linear theory of orographic precipitation (ADLOP). The schemes are tested over the Attica region of Greece using three high-impact precipitation events representing different synoptic weather regimes and verified against high-resolution weather radar observations. Forecast performance is assessed using continuous, categorical, and neighborhood-based spatial verification metrics. Results show that the Control performs competitively for light precipitation and at larger neighborhood sizes in localized events. The inclusion of diffusion in the AD scheme generally reduces random errors. In this limited three-case sample, aggregated results show that ADLOP reduces systematic bias, with reductions reaching approximately 33% at longer lead times and showing higher detection scores. However, its added value is strongly dependent on the terrain-influenced precipitation regime and may be accompanied by increased error at longer lead times. Overall, the benefits of incorporating diffusion and simplified linear orographic forcing depend on precipitation regime, lead time, and verification metric; therefore, the results should be interpreted as diagnostic case-study evidence rather than as a general assessment of ADLOP performance. Full article

The Western Balkans (Serbia, Croatia, Bosnia and Herzegovina, and Montenegro) occupy a transitional climatic position between the Mediterranean hotspot and the continental Balkan interior within Southeast Europe, yet multi-country, multi-model, station-resolved assessments of regional aridification remain scarce. We combined quality-controlled monthly temperature and precipitation records from 100 stations across Serbia, Croatia, Bosnia and Herzegovina, and Montenegro (1961–2020) with bias-corrected projections from a five-member CMIP6 ensemble (EC-Earth3, MPI-ESM1-2-HR, CNRM-CM6-1, MRI-ESM2-0, IPSL-CM6A-LR) under four SSP scenarios to 2100 and benchmarked these projections against five short-term forecasting baselines on a held-out 2019–2020 period. Aridity was quantified using the Ellenberg Climate Quotient (EQ) and De Martonne Index. Results: Ninety-eight of 100 stations showed significant warming (1961–2020, p < 0.05), and 26 showed significant aridification. Ensemble mean end-of-century EQ change ranged from −6.9% (SSP1-2.6) to +44.6% (SSP5-8.5, Montenegro), with the largest absolute increases in the Pannonian lowlands; inter-model uncertainty exceeded inter-scenario uncertainty by roughly a factor of two. Deep learning forecasters (TFT, N-HiTS) outperformed bias-corrected CMIP6 output for short-term, station-scale temperature forecasting, while a simple climatological baseline remained competitive for precipitation. CMIP6 projections and AI forecasting are complementary: multi-model ensembles remain indispensable for long-term, scenario-conditioned planning, while AI offers superior near-term predictive skill for operational decisions. Full article

Modern power transmission grids are increasingly operated under volatile load, variable generation, and near-limit line-flow conditions. In such environments, deterministic line-flow forecasting is insufficient for operational decision support because operators require calibrated risk probabilities, uncertainty intervals, and reliable early warning signals. This article proposes B-TGPRF, a Bayesian Temporal Graph Probabilistic Risk Forecaster for calibrated overload-risk forecasting in power transmission grids. The proposed framework is positioned as a hybrid probabilistic graph-temporal forecasting system rather than as a new end-to-end graph neural network; its novelty lies in the leakage-controlled sequential combination of temporal forecasting, electrical graph descriptors, calibrated Bayesian risk estimation, residual correction, and compact interval uncertainty assessment. The model integrates temporal line-flow, load, and generation features with graph-topological descriptors, operating-regime indicators, residual correction, conformal interval estimation, probability calibration, and a Bayesian risk layer. A leakage-controlled data preparation pipeline was built using an open large-scale benchmark for machine learning applications in transmission grids. The final modeling dataset contains more than 3.48 million observations, 100 selected critical lines, train-only risk thresholds, and chronological train, validation, test, and external-like scenario splits. B-TGPRF was compared with persistence baselines, linear models, Bayesian baselines, tree ensembles, boosting models, neural temporal models, and compact state-of-the-art-style temporal and graph-temporal architectures. On the strict external-like test, B-TGPRF achieved MAE = 0.3650, RMSE = 0.5478, R² = 0.9962, Brier Score = 0.0147, and ECE = 0.0064. The results show that the proposed model provides a strong overall balance between line-flow forecasting accuracy, calibrated risk estimation, compact interval prediction, and low false-positive risk-signaling, while boosting models remain highly competitive for pure risk-class detection. Full article

High-quality renewable energy (RE) scenario generation is essential for secure, reliable, and economic power system operation under high RE penetration. To address the limitations of existing methods in preserving scenario fidelity, diversity, spatiotemporal dependence, and engineering consistency, this paper proposes a three-stage framework that combines a variational autoencoder (VAE) with a conditional latent diffusion model (CLDM), hereafter referred to as VAE-CLDM, for day-ahead renewable energy scenario generation. First, multi-source features are constructed by integrating renewable power outputs, meteorological variables, temporal lag information, and spatial correlation characteristics among wind farms and photovoltaic stations. Then, VAE compresses the high-dimensional features into a compact latent space while retaining key statistical and spatiotemporal information. Based on this latent representation, a CLDM generates realistic scenarios by progressively denoising random noise under meteorological conditions. A spatiotemporal feature modeling strategy is incorporated to better represent temporal fluctuations and inter-site correlations, while a diversity regulation factor is selected on the validation set to balance scenario fidelity and tail-event coverage. Finally, the generated scenarios are reconstructed into the physical space and checked using output-bound and ramp-consistency correction to improve their practical usability. Results on the open dataset released by the Chinese State Grid Renewable Energy Generation Forecasting Competition show that the proposed framework achieves the lowest root mean square error (RMSE), mean absolute error (MAE), and maximum mean discrepancy (MMD) among the tested models on the training-statistics-standardized renewable-power benchmark, with RMSE of $0.3013\pm 0.0022$, MAE of $0.3636\pm 0.0010$, and MMD of $0.04269\pm 0.00040$. After post-correction, lower- and upper-bound violations are reduced to 0.00%, and the ramp-violation rate is reduced to 0.08%, indicating that the proposed VAE-CLDM can provide useful scenario inputs for day-ahead dispatch and risk assessment in renewable-dominated power systems. Full article

Short-term traffic flow forecasting becomes especially difficult when incomplete observations, within-window frequency variation, and state-dependent sensor relations occur together. Missing readings can affect both node features and the spatial dependencies inferred from them, yet these issues are commonly modeled separately. We therefore propose RIFT-STGNN, a Robust Interleaved Frequency–Trend Spatio-Temporal Graph Neural Network for multi-step traffic flow forecasting. RIFT-STGNN follows a coordinated information flow: observation status is retained during temporal–frequency encoding, the frequency representation supports both node features and graph construction, and four graph sources are fused before each Graph-GRU update. Trend-aware temporal attention then produces direct multi-step forecasts. Experiments on PEMS03, PEMS04, PEMS07, and PEMS08 show competitive numerical performance relative to selected literature-reported baselines under the 12-step setting. On PEMS04 and PEMS08, the three-run mean MAE values are 17.73 and 13.14, respectively. These values are numerically 4.63% and 9.00% lower than the corresponding literature-reported sAMDGCN values. Component ablations, state-dependent graph analysis, and controlled missing-rate experiments support the roles of dynamic graph learning, within-window frequency encoding, and mask-aware input handling under the evaluated settings. Full article

This work presents a CUDA-accelerated methodology for training multiple neural networks in parallel using population-based metaheuristics. The goal is to obtain fast and accurate short-term energy-forecasting models for time-sensitive building-management applications. We evaluate five metaheuristic optimizers and their memetic variants, for which a local-search stage based on the $ADAM$ optimizer is incorporated. The models are assessed on eleven real-world energy-consumption time series using training time, root mean squared error (RMSE), mean absolute error (MAE), and normalized RMSE (NRMSE). The results show that the proposed memetic approaches are competitive under strict training-time budgets, although unconstrained $ADAM$ remains a strong overall reference. Full article

This study aims to accurately predict students’ academic performance trajectories for university entrance examinations by proposing a machine learning framework that explicitly accounts for missing data and uncertainty. Mock examination data are characterized by substantial missing values due to heterogeneous participation in exams, as well as inherent randomness caused by variations in test content and examinee conditions. Conventional single-value imputation methods cannot adequately reconstruct the missing values arising from such heterogeneous participation without introducing strong bias, and existing educational prediction models based on deterministic formulations do not account for the inherent randomness and uncertainty in examination scores, thereby limiting the reliability of their forecasts. To address these challenges, we employ GP-VAE and SAITS, state-of-the-art methods for time-series imputation, to reconstruct incomplete mock examination data. Furthermore, we develop a Bayesian Neural Network (BayesNN) to predict future academic performance while explicitly modeling uncertainty. By integrating temporally aware imputation with probabilistic prediction, the proposed framework aims to provide more accurate and reliable performance forecasts than existing approaches. We evaluate the effectiveness of the proposed method through comparative experiments involving various combinations of imputation techniques and prediction models. Experimental results demonstrate that the proposed framework achieves competitive predictive accuracy: the combination of deep imputation methods and BayesNN yields the lowest average estimation error of 15.98 points, compared with 16.75 points for the conventional combination of mean imputation and linear regression. The contribution of this study does not lie in proposing a new deep learning model itself, but rather in systematically comparing combinations of time-series imputation methods and uncertainty-aware prediction models using real-world mock examination sequence data with missing values, thereby providing effective design guidelines for educational data analysis. Full article

Infectious disease outbreaks strain medical resources and public-health systems, making accurate multi-horizon, multi-regional forecasting essential for early warning and resource allocation. Existing approaches face a trade-off: data-driven deep models can capture nonlinear spatiotemporal patterns but often overlook transmission mechanisms, whereas mechanistic models are interpretable but limited in modeling complex regional dependencies. To address this challenge, we propose Epidemic Spatial–Temporal Large Language Model (EpiSTLLM), which is a graph-structure-enhanced large language model for multi-regional infectious disease forecasting. EpiSTLLM injects regional adjacency information into Transformer-based representation learning to capture cross-regional transmission structure and long-term temporal dependencies. A temporal-gated cross-attention module generates horizon-specific latent transmission and recovery parameters, while a latent-space SIR-inspired propagation mechanism with a residual correction branch enables stable multi-horizon forecasting without requiring fully observed compartmental states. Experiments on the FluView state-level influenza-like illness dataset and NHSN state-level influenza hospitalization dataset show that EpiSTLLM achieves the best or highly competitive performance in most evaluation settings against statistical, deep learning, graph-based, and mechanism-guided baselines across 4-, 8-, and 12-week horizons. For example, EpiSTLLM reduces MAE and RMSE values by $9.5\%$ and $6.2\%$ at $H=4$ on FluView, and by $12.0\%$ and $13.1\%$ at $H=8$ on NHSN compared with the strongest baselines, respectively. Full article

This study presents a spatially aware machine learning model for probabilistic groundwater potential assessment using multimodal geospatial data derived from topography, hydrotopography, climate, soil, land use, Sentinel-1 SAR, and water balance variables. The model incorporates spatially consistent data partitioning, leakage-controlled model development, and probabilistic forecasting to improve the robustness and transferability of groundwater potential assessment. A total of 2402 spatial observations, including 601 groundwater-related locations and 1801 spatially filtered pseudo-absence samples, were used to evaluate eleven machine learning and deep learning models. The proposed hybrid framework achieved the best overall validation results with ROC-AUC of 0.9319, PR-AUC of 0.8363, F1-measure of 0.8226, balanced accuracy of 0.8847, and MCC of 0.7613. Independent spatial block testing further demonstrated strong generalization ability, showing ROC-AUC of 0.9535, PR-AUC of 0.9002, F1-measure of 0.7983, balanced accuracy of 0.8594, and MCC of 0.7336. Comparative experiments demonstrated that the proposed framework remains competitive with state-of-the-art machine learning and deep learning approaches while providing robust probabilistic estimates in spatially separated validation. The resulting groundwater potential maps identify areas with environmental conditions similar to known groundwater observations and provide a reliable basis for prioritizing hydrogeological studies, groundwater exploration, and regional water resources planning in data-poor settings. Full article

Accurate forecasting of country-level temperature anomalies is increasingly important for climate monitoring, policy planning, and environmental risk assessment. However, the trade-off between predictive performance, model complexity, and computational cost remains insufficiently explored, particularly across multiple countries using compact and interpretable feature representations. This study presents a comprehensive comparative evaluation of eight forecasting approaches for annual temperature anomaly prediction using country-level observations from the FAOSTAT Temperature Change dataset. The evaluated methods comprise a Persistence baseline, Ordinary Least Squares (OLS), Ridge regression, Support Vector Regression (SVR), Random Forest, a multilayer perceptron (MLP), and the classical time-series models ARIMA and ETS. Annual temperature anomalies were modeled using lagged observations, a temporal trend, and a trailing moving average under a temporally ordered 80/20 train–test split. Model performance was assessed using RMSE, MAE, $R^2$, per-country win-rate, computational runtime, and pairwise statistical comparisons based on the Wilcoxon signed-rank test with Holm correction. Hyperparameters were optimized through expanding-window temporal cross-validation, and an ablation study was conducted to quantify feature contributions. Results indicate that the ETS model achieved the best overall predictive performance, obtaining the lowest median RMSE (0.3388 °C), the lowest MAE (0.2792 °C), and the highest per-country win-rate (40.07%). ARIMA provided competitive forecasting accuracy but incurred substantially higher computational cost, whereas OLS and Ridge offered an attractive compromise between predictive performance, robustness, interpretability, and computational efficiency. In contrast, the more flexible machine learning models (SVR, Random Forest, and MLP) did not consistently outperform the simpler approaches despite their higher complexity. Overall, the results demonstrate that classical statistical forecasting methods remain highly competitive for annual country-level temperature anomaly prediction and that increasing model complexity does not necessarily translate into improved predictive performance. Full article

Accurate wind vector prediction is essential for renewable energy utilization and power system stability, yet existing methods struggle to jointly model local dynamics, global structures, and temporal robustness. To address this limitation, an Adaptive Multi-Branch Heterogeneous Fusion Wind Prediction Network (AMBHFN) is proposed. Local dynamic, global structural, and temporal robustness modeling are assigned to dedicated heterogeneous branches, whose outputs are coordinated through the Adaptive Multi-Branch Prediction Collaboration Mechanism (AMBPC). Multi-source meteorological variables and terrain information are used for local dynamic modeling, while global spatiotemporal structures are captured by a 3D U-shaped fully convolutional branch and temporal robustness is enhanced by an iTransformer-based multi-agent branch with graph convolution. Experiments on ERA5 data show that AMBHFN outperforms eight retrained baselines over the 0–23 h forecast horizon, with an average error reduction of more than 12%. At the first forecast step, the root mean square error (RMSE) and mean absolute error (MAE) are 0.33 m/s and 0.25 m/s, respectively. Under the strict 22.5° threshold, wind direction forecast accuracy (WDFA) reaches 97.72% at 0 h and 78.06% at 6 h. Fine-tuning in two target regions reduces the 13–23 h RMSE to 1.54 and 1.96. Statistical tests confirm significant improvements over MFWPN, and ablation studies verify the complementarity of the three branches. With 128 giga floating-point operations (GFLOPs) and a 22 ms per-sample forward inference time, AMBHFN achieves a competitive balance among accuracy, stability, and efficiency. Full article

Sub-Saharan Africa is rich in agricultural biomass, yet its biogas utilization is far below its potential. Most earlier studies failed to identify the nonlinear, multi-factor relationships that shape real national biogas yields and fully clarify this imbalance. This study constructs a 2007–2023 panel dataset for ten sub-Saharan African countries, merging agricultural output, socioeconomic, and infrastructure metrics. Gradient Boosting model and SHapley Additive exPlanations (SHAP) analysis are applied for empirical evaluation. SHAP analysis confirms that charcoal consumption yields the largest contribution to biogas production, with a mean absolute SHAP value of 1.018. The correlation between the two variables is negative under the threshold and becomes positive beyond this critical level. Urbanization has an inverted U-shaped correlation with biogas output, and the marginal contributions of predictors vary substantially across sampled countries. Instead, fragile supply chains, rural labor loss, and fierce competition in clean energy markets curb local biogas production. Forecasts show that regional biogas output will continue to fall until 2030. Targeted national policies matching each country’s core influencing factors are therefore urgently required. Full article