Search Results (27,398)

Rare event detection in large tabular data remains a computationally challenging problem due to class imbalance, heterogeneous feature distributions, and unstable thresholds. Traditional machine learning approaches based on individual models and fixed thresholds often exhibit limited robustness and reproducibility in such settings. This paper proposes a hybrid ensemble framework for rare event detection that integrates heterogeneous machine learning models through threshold-aware probabilistic aggregation. The framework combines gradient-boosted decision trees, regularized linear models, and neural networks, leveraging their complementary inductive biases. To ensure reproducibility and robust performance evaluation under severe class imbalance, a leaky-controlled evaluation protocol is employed, including rootwise summation, probability calibration, and validation-based threshold optimization. The proposed approach is evaluated on a large tabular dataset containing approximately 50,000 observations. Experimental results demonstrate improved rare event detection and robust generalization performance compared to individual baseline models. Explainability is achieved through Shapley Additive Explanations (SHAP)-based attribution analysis and clustering in the explanation space, enabling transparent analysis of ensemble decision-making behavior. The proposed framework represents a general-purpose computational solution for rare event detection and can be applied to a wide range of data-driven decision-making and anomaly detection problems. Full article

To address stability issues induced by dynamic fracture of weak interlayers in sandstone–mudstone interbedded slopes during blasting excavation, this study investigates the Qingnian Hub diversion channel project of the Ping-Lu Canal through an integrated methodology combining field blasting tests, laboratory dynamic rock experiments, and numerical simulation validation. Field monitoring captured slope dynamic responses, while ultrasonic testing and Split Hopkinson Pressure Bar (SHPB) dynamic splitting tests determined rock mass mechanical parameters. A high-fidelity 3D numerical model developed in ANSYS/LS-DYNA was validated against experimental data, demonstrating reliability with relative errors in peak particle velocity (PPV) below 20% at most monitoring points. Results reveal that increasing interlayer dip angle reduces fracture length along the lower interface while causing internal oblique cracks to initially lengthen and then shorten, with optimal oblique crack development observed at 10–15°. Conversely, greater interlayer spacing first decreases and then stabilizes lower-interface fracture length, whereas oblique crack length peaks at 4.8 m for a 4 m spacing. Based on 25 parametric simulations, a safety criterion using crack-initiation vibration velocity was established, yielding a predictive model dependent on dip angle and spacing. The derived criterion defines a critical vibration velocity range of 5.6–10.0 cm/s for the studied slope configurations. Compared to existing empirical guidelines that rely solely on peak particle velocity, the proposed criterion innovatively incorporates the controlling influence of geological stratigraphic geometry. This study provides theoretical and practical guidance for optimizing blasting parameters and ensuring slope stability in similar engineering contexts. Full article

Monitoring environmental changes over time requires images with extensive historical depth. However, high spatial resolution images often lack such depth. This study investigates the impact of spatial resolution on image classification. Thus, Landsat 8 and Sentinel-2 images acquired between October and December 2020 were processed and classified using Random Forest regression on Google Earth Engine (GEE). This method allows for continuous land cover maps, required for robust assessment of land cover dynamics in patchy landscapes. A total of 1719 training samples were collected from the Collect Earth Online (CEO) platform to train the model. In addition to the spectral bands, vegetation indices were considered to optimize classification results. The study revealed statistical differences in land cover areas estimated by the two sensors. These differences are statistically significant at p < 0.001, although they are small. Validation results showed that the RMSE from Sentinel-2 is slightly lower than that from Landsat 8, with this difference significant at p < 0.05. Therefore, spatial resolution influences the accuracy of image classification. Nevertheless, given the observed differences between the two sensors, which ranged from 0.03% to 3.94% across land covers, Landsat imagery remains suitable for producing reliable land cover maps in heterogeneous landscapes. Full article

This study provides computational insights into the flexural strengthening of reinforced concrete (RC) T-beams in the negative moment region using steel-reinforced polymer cement mortar (PCM) overlays. A validated three-dimensional nonlinear finite element (FE) model was developed using the Advanced Tool for Engineering Nonlinear Analysis (ATENA) software (version 2023.0.0.22492) to simulate the behavior of beams retrofitted with 40 mm thick PCM layers embedded with 13 mm and 16 mm deformed bars. Model validation was performed against previously published experimental results reported by the authors, demonstrating excellent agreement, with normalized mean square error (NMSE) values expressed as fractions between 0.0001 and 0.0022, and experimental-to-numerical ultimate load ratios ranging from 0.99 to 1.01. Parametric analyses were then conducted to investigate the influence of key variables, concrete compressive strength, PCM overlay thickness, and longitudinal reinforcement ratio on the global flexural performance. The results revealed that increasing the overlay thickness raised the ultimate load capacity by up to 15.4% and improved energy absorption by 43%. Enhancing concrete strength led to gains of up to 12.5% in load capacity and 15.8% in stiffness. Variations in reinforcement ratio had the most significant impact, increasing peak load by up to a factor of 2.02 and improving energy absorption by up to a factor of 1.49. Despite these improvements, reductions in ductility were observed across all strengthening configurations, underscoring a strength–deformability trade-off critical for seismic applications. These findings affirm the efficacy of steel-reinforced PCM overlays and provide design-oriented insights for optimizing negative moment retrofitting strategies in RC bridge girders and continuous beam systems. Full article

Constructed wetlands (CWs) have emerged as effective nature-based solutions (NbS) for the treatment of industrial dairy wastewater (DWW), which is characterized by high organic loads, elevated nutrient concentrations, and pronounced operational variability. Despite increasing implementation, quantitative engineering evidence supporting design optimization and scalability remains fragmented. Herein, we present a semi-quantitative synthesis of CW performance for DWW treatment, explicitly linking hydraulic and operational parameters with pollutant removal efficiencies. A systematic review of 38 peer-reviewed studies published between 1995 and 2025 was conducted in accordance with PRISMA 2020 guidelines. Treatment performance was normalized and evaluated as a function of hydraulic retention time (HRT), organic loading rate (OLR), system configuration, and climatic context. The results demonstrate that hybrid CWs combining vertical and horizontal subsurface flow most frequently achieved COD and BOD₅ removal efficiencies exceeding 90% when operated within an observed operating envelope, typically including HRT ranges of 4–8 h (VSSF; n = 4) and 3–7 days (HSSF; n = 14), and OLR values below 30 g COD m⁻² d⁻¹ (n = 7, among studies reporting OLR). Operation outside this operating envelope was generally associated with reduced treatment stability and an increased likelihood of operational constraints (e.g., clogging). Substrate porosity, vegetation diversity, and climate further modulated long-term performance and system resilience. Based on the consolidated evidence, this review suggests transferable operational design envelopes and configuration-specific implementation pathways that translate empirical findings into practical engineering guidance, supporting the scalable adoption of CWs as low-energy NbS for decentralized and sustainable DWW management. Full article

This study’s aim was to clarify the regulatory roles of the MKI67 rs11016073 and APOB rs1367117 polymorphisms in the relationship between blood Pb levels and cancer risk. Blood Pb concentrations were measured using inductively coupled plasma mass spectrometry, and genotyping was performed by real-time PCR with TaqMan probes. Cancer incidence was assessed during a mean follow-up of six years and two months. During follow-up, 210 incident cancers were diagnosed among 2782 women. Pb exposure was categorized into quartiles (Q1: <9.44 µg/L; Q2: 9.44–12.58 µg/L; Q3: 12.59–17.16 µg/L; Q4: >17.16 µg µg/L). The association between Pb levels and cancer risk was strongly genotype dependent. Women carrying APOB non-GG and MKI67 non-AA genotypes exhibited the lowest breast cancer risk at the highest Pb levels (Q4), whereas carriers of APOB GG and MKI67 AA showed the lowest risk at the lowest Pb levels (Q1). Age-stratified analyses further demonstrated genotype-specific differences in optimal Pb exposure ranges, particularly for breast cancer. Cancer risk associated with Pb exposure is not uniform but depends on genetic background. These findings identify genotype-specific optimal blood Pb levels and suggest that incorporation of MKI67 and APOB genotyping may improve risk stratification and interpretation of non-linear Pb–cancer associations. Full article

L-Ergothioneine (L-EGT), a naturally occurring thiol compound abundant mainly in edible fungi, is increasingly regarded as a potentially beneficial bioactive constituent. However, population-level exposure data remain limited. This study aimed to estimate background dietary exposure to L-EGT among Chinese residents, describe its distribution across population subgroups and regions, identify major food contributors, and characterize the risk using a margin of exposure (MOE) approach. Individual body-weight-normalized L-EGT intakes were estimated from published food concentration data and three-day dietary recalls of 42,218 participants. MOEs were calculated using a no observed adverse effect level (NOAEL) of 800 mg/kg bw/d obtained from subchronic toxicity studies. The mean dietary exposure to L-EGT was 0.043 mg/kg bw/d (MOE = 18,605) in the general population and 0.174 mg/kg bw/d (MOE = 4598) among consumers, with 95th percentile exposures of 0.244 mg/kg bw/d (MOE = 3279) and 0.644 mg/kg bw/d (MOE = 1242), respectively. MOE values were consistently above the safety threshold of 300 across all subgroups, with less than 0.3% of the total population and 1.3% of consumers aged 3–6 years falling below this value. These results indicate that current natural dietary exposure to L-EGT in China is low and does not raise safety concerns and provide essential baseline data for future studies on its health effects, optimal intake ranges, and long-term safety. Full article

Automated electrocardiogram (ECG) arrhythmia classification remains challenging due to signal noise, inter-patient variability, and limited annotated data, which constrain the generalization of supervised learning approaches. This study presents a self-supervised ECG representation learning framework that combines contrastive pretraining with ensemble-based supervised classification. A signal-to-noise ratio criterion is applied during self-supervised pretraining to stabilize contrastive optimization, while all extracted ECG beats, including noisy segments, are retained during downstream evaluation. The learned representations are classified using a hybrid ensemble composed of convolutional encoders and tree-based models. Model evaluation follows strict patient-level partitioning with stratified 10-fold cross-validation and bootstrap-based uncertainty estimation on a held-out test set. Under this evaluation protocol, the framework achieved high beat-level performance on curated datasets (internal and external). Class-wise performance shows precision and recall values between 0.99 and 0.999 across normal, supraventricular, ventricular, fusion, and paced beat categories. External validation is conducted on independent ECG cohorts, including PTB-XL, Chapman–Shaoxing, and INCART 12-lead datasets. On these datasets, the hybrid model attains macro-F₁ scores ranging from 0.91 to 0.94, compared with standalone convolutional and handcrafted feature-based Random Forest classifiers evaluated under identical conditions. These results characterize the behavior of the proposed representation learning framework across heterogeneous patient populations and recording configurations. Full article

This paper presents a systematic optimization methodology for reluctance magnetic gears (RMGs) using genetic algorithms to maximize peak torque within fixed volume constraints. A multi-run parametric study with population sizes ranging from 50 to 150 individuals demonstrates robust convergence toward remarkably similar geometric proportions, with a best torque of 430.15 Nm and minimal variation (0.65%) across different configurations. The optimization redistributes radial dimensions, achieving a 95% torque improvement over uniform distribution by significantly increasing inner rotor teeth thickness by 85% while reducing modulator thickness by 74%. The most significant outcome is the emergence of a consistent design trend across all independent runs, providing general guidelines on how the RMG internal geometry should be arranged to maximize peak torque. Full article

Background/Objectives: Neck disorders encompass a range of discomforts impacting a person’s quality of life. Traditional diagnostic methods, such as physical tests and imaging techniques, rely heavily on clinician expertise, leading to potential variability in assessments. While ultrasound imaging is commonly used, the application of machine learning models to assess neck disorders, particularly fascial abnormalities, remains limited. This study seeks to fill this gap by developing a machine learning model using ultrasound images to provide accurate and efficient support for diagnosing neck disorders. Methods: Due to limited availability of labeled ultrasound data for neck disorders, developing robust and generalizable models remains a challenge. In this study, a neck disorder assessment system was developed using ultrasound images collected from 184 patients by employing various machine learning algorithms. To address data scarcity and improve model generalizability, an approach utilizing EfficientNet with transfer learning was introduced and thoroughly assessed using the trained model on a completely clean test dataset, ensuring the robustness of the solution. The model was trained using 5-fold cross-validation with the respective weight of each class and AdamW as the optimizer. Results: The results showed promising performance, with the deep fascia fuzzy texture and deep fascia and myofascial adhesion at lower cervical regions demonstrating the highest weighted average F1-scores of 76% and 81%, respectively. The macro averages reflected similar performance, at 74% and 78%, respectively, indicating consistent class-wise accuracy for these regions. Conclusions: The proposed model demonstrated robust classification performance for neck disorder assessment, particularly in evaluating the lower cervical region. This approach has the potential to support clinical decision-making by providing consistent, efficient, and accurate diagnostic assistance. Further refinement and validation across diverse clinical settings will be critical to enhance its real-world applicability. Full article

Air-cooled heat sinks remain a practical and cost-effective solution for thermal management in high power-density electronic systems. This study investigates the thermal–hydraulic performance of a plate pin-fin heat sink operating under forced convection, with emphasis on the coupled interaction between heat-transfer enhancement and pressure-drop penalty. The proposed hybrid configuration combines the low flow resistance of plate fins with the wake-induced mixing characteristics of pin-fin elements, thereby modifying boundary-layer development and flow structures within the fin channels. This review comprehensively analyzes existing experimental measurements across a range of Reynolds numbers to evaluate the average Nusselt number, thermal resistance, and friction factor. The results demonstrate that the inclusion of pin elements significantly enhances convective heat transfer through increased flow disruption and vortex formation, while incurring a moderate increase in pressure loss relative to conventional plate-fin designs. In addition, flow visualization and temperature mapping reveal improved heat transfer uniformity along the streamwise direction, particularly at intermediate Reynolds numbers where transition effects become pronounced. Empirical correlations were developed to relate the Nusselt number and friction factor to Reynolds number and key geometric ratios, providing predictive capability for thermo-hydraulic performance assessment. The findings indicate that fin-scale geometric optimization plays a dominant role in achieving improved overall performance and that the plate pin-fin configuration offers a favorable trade-off between heat-transfer augmentation and hydraulic efficiency for forced-convection electronic cooling applications. Full article

To support carbon stock assessment and ecological restoration under the “Carbon Neutrality” objective, this paper developed a high-precision vegetation biomass model for expressway corridors in Shanxi Province, China, by integrating Unmanned Aerial Vehicle(UAV) technology and the random forest algorithm. Based on climatic zoning and DEM data, 70 sample plots representing diverse vegetation and topography were selected. LiDAR point clouds and multispectral data were spatially connected using the BallTree algorithm, achieving an average matching rate of 73.98–82.01%. A joint biomass model incorporating tree height and crown width was constructed with spatial cross-validation. The results indicate that the model substantially outperformed single-factor models, with R² values ranging from 0.839 to 0.934 (highest in the Hengshan–Wutaishan forest area). Accuracy was higher in forest-dominated zones but lower in areas with significant human disturbance. A representative sample library was established for model optimization. This paper provides a robust technical framework for biomass monitoring across comparable Northern Hemisphere latitudes, thereby supporting sustainable green transport development. Full article

Global water scarcity is intensifying, and agriculture remains the main consumer of freshwater. Many studies have assessed agricultural water productivity (WP) in major farming regions. While previous studies have mainly assessed overall efficiency or single crops, crop-specific dynamics within double-cropping systems remain insufficiently understood. This study quantifies the spatial patterns and stage-wise changes in winter wheat and summer maize WP in the North China Plain based on five representative years (2000, 2005, 2010, 2015, and 2019) and examines their climatic and anthropogenic drivers. The Optimal Parameter-Based Geographical Detector (OPGD) model was used to assess the explanatory power of influencing factors, and the Multi-scale Geographically Weighted Regression (MGWR) model was applied to capture spatially heterogeneous relationships. Wheat WP ranged from 0.56 to 1.30 kg m⁻³ and showed a significant increasing trend, whereas maize WP ranged from 0.89 to 1.72 kg m⁻³. Both climatic and anthropogenic factors exhibited pronounced spatial heterogeneity. Beijing and Tianjin were classified as anthropogenic-dominated zones, while several cities in Henan displayed crop-specific dominant drivers. Fertilizer application was negatively associated with WP in multiple regions, indicating declining input efficiency under intensive management. These findings support irrigation zoning and differentiated water allocation strategies, contributing to sustainable intensification and progress toward water-related (SDG 6) and food security (SDG 2) goals in intensive double-cropping regions. Full article

Scientific concern regarding the widespread occurrence of micropollutants (MPs) in aquatic environments has been growing. Background/Objectives: Since conventional wastewater and drinking water (DW) treatment plants are generally unable to completely remove MPs, their presence in DW may occur, potentially posing adverse effects on public health. Highly sensitive analytical methods are crucial, as MPs may occur at very low concentrations in DW, usually at ng L⁻¹ levels. Methods: An offline solid-phase extraction ultra-high performance liquid-chromatography coupled to tandem mass-spectrometry (SPE-UHPLC-MS/MS) method was optimized and validated for the determination of 23 MPs in DW, including 12 pharmaceuticals, 9 pesticides, and 2 UV-filters, listed in the 2 most recent European Union (EU) Decisions (2022/1307 and 2025/439) for surface water monitoring, and in the revised EU Urban Wastewater Treatment Directive (2024/3019). The validated method was applied to 50 DW samples collected across Portugal. Results: The optimized SPE-UHPLC-MS/MS method showed high analytical sensitivity, achieving method detection limits below 1.50 ng L⁻¹. Up to 3 MPs were detected per sample, with quantifiable concentrations of each ranging from 0.28 to 98.8 ng L⁻¹. However, benzotriazole and dimoxystrobin exceeded the upper limits of their calibration curves (i.e., concentrations higher than 133 and 117 ng L⁻¹, respectively) in one and 3 of the collected samples, respectively. Considering all analyzed samples, 4 (fluconazole, irbesartan, dimoxystrobin, and benzotriazole) of the 23 target compounds were detected. Hazard quotient values for all detected MPs were below 0.2. Conclusions: The validated SPE-UHPLC-MS/MS method is suitable for the sensitive determination of MPs in DW. Some MPs were detected, with concentrations indicating no expected human health risks under the conditions evaluated. Further monitoring campaigns should be conducted in the future, with compounds exceeding the limits of the calibration curves requiring special attention. Full article

This paper presents a novel optimization algorithm for electric vehicle (EV) aggregators aiming to maximize net revenue in demand response markets. Aggregated EV charging stations are modeled as a battery with time-varying capacity, enabling participation in these markets. Due to uncertainties in EV plug-in duration and energy demand, it is challenging for aggregators to fulfill bid capacities in real-time (RT). To address this, EV users specify minimum acceptable service levels, allowing aggregators to optimize both charging timing and energy demand in RT. The model is composed of two layers: (1) a Day-Ahead (DA) optimizer that determines optimal EV scheduling and DA demand response market bidding, and (2) a two-stage RT optimizer that fine-tunes the charging schedule using real-time flexibility to mitigate forecast errors. The RT optimizer leverages Model Predictive Control (MPC) in a two-stage structure to address the problem’s non-convexity, which arises from two coupled unknowns: the charging time and the charging energy demand. In the first stage, it determines a cost-optimal charging schedule that ensures full service levels. In the second stage, it optimizes the charging energy demand within a feasible range, bounded above by the first-stage trajectory and below by user-defined minimum service levels, to maximize demand response market revenue. A realistic baseline and a penalty term are integrated into the demand response market revenue term of the cost function to more accurately reflect real-world conditions. Simulation results demonstrate that the proposed method yields a net economic profit at least five times higher than that of immediate (or `dumb’) charging. During one month of simulations, the aggregator achieves revenue equivalent to $0.21 per kWh of demand reduction under forecast uncertainty, totaling $3441. Full article