Search Results (4,105)

Background/Objective: Activated coagulation factor XII (FXIIa) is a component of the contact activation pathway and a pharmacologically relevant target in contact-system-associated processes. In this study, scaffold-aware and interpretable machine-learning QSAR models were developed for human FXIIa activity. Methods: Bioactivity records for the human single protein target CHEMBL2821 were retrieved from ChEMBL release 37. Modeling was restricted to exact IC50 measurements from assays explicitly referring to FXIIa, Factor XIIa, or activated Factor XII. Median-consolidated pIC50 values and two-dimensional Mordred descriptors were evaluated using leakage-safe preprocessing, scaffold-disjoint validation, Y-randomization, applicability domain analysis, structural similarity auditing, and SHAP interpretation. Results: The regression dataset comprised 424 compounds and 166 Bemis–Murcko scaffolds in this study. The Gradient Boosting regressor achieved R² = 0.7560, RMSE = 0.6924, and MAE = 0.4965 on the locked scaffold-disjoint test set (n = 85); across 50 repeated scaffold partitions, the mean R² was 0.6892 ± 0.1515. The classification model achieved ROC-AUC = 0.9453, PR-AUC = 0.9807, balanced accuracy = 0.7561, and MCC = 0.5972 (n = 73). Y-randomization supported nonrandom predictive signals (empirical p = 0.0099). Conclusions: The models support computational prioritization within the represented FXIIa chemical domain, while prospective evaluation of independently generated compounds remains necessary. Full article

Physical sensor deployments in ground-based environmental monitoring networks are frequently constrained by high installation costs, hardware failures, and limited spatial coverage, resulting in incomplete observational datasets and degraded sensing capacity across monitoring stations. Data-driven virtual sensing offers a cost-effective alternative by estimating target environmental parameters through machine learning models trained on correlated sensor measurements, reducing dependency on dense physical infrastructure. This paper presents VS-DCFF, an applied virtual sensing framework for dual-target estimation of near-surface air temperature and relative humidity from ground-based sensor network data. VS-DCFF integrates: (i) a deterministic pipeline applying temporal encoding, rolling-window statistics, and mutual information-based feature selection to capture a linear trend/seasonal component and to select features predictive of the residual signal; (ii) a probabilistic pipeline employing a Gaussian copula model to generate statistically consistent synthetic residual samples preserving inter-variable dependencies; and (iii) an early feature-level fusion strategy feeding a copula-augmented XGBoost residual-boosting stage, whose output is combined with the linear trend component for the final prediction. Under a strict chronological evaluation protocol, VS-DCFF is benchmarked against persistence, linear and ensemble regression baselines, and a same-protocol re-implementation of the statistical core of the VSG-SGL framework, and achieves near-surface air temperature $\mathrm{RMSE}=0.7791{}^{\circ }$C, $R^2=0.9735$, and relative humidity $\mathrm{RMSE}=3.7747\%$, $R^2=0.9701$, outperforming all tested baselines. The framework is further validated through leave-one-station-out spatial generalization, robustness evaluation under simulated sensor faults and target-history loss, copula-variant and synthetic-data fidelity diagnostics, and a lightweight edge-deployment ablation. All findings, including cases where tested extensions such as spatial context features did not yield a robust improvement, are reported transparently. Results indicate that the proposed architecture provides a computationally efficient, extensively validated approach to dual-target environmental virtual sensing under realistic deployment conditions. Full article

Reliable clinical decision support systems require not only high predictive accuracy but also trustworthy probability estimates and robust uncertainty quantification. However, most medical artificial intelligence (AI) studies primarily emphasize discrimination performance while overlooking systematic reliability evaluation. This study proposes a reliability-aware evaluation framework for Alzheimer’s disease detection that integrates discrimination analysis, statistical validation, probability calibration, uncertainty quantification, robustness assessment, and clinical decision analysis within a unified pipeline. Multiple machine learning classifiers and ensemble configurations were evaluated using repeated stratified cross-validation and assessed through discrimination and calibration metrics. Support Vector Machine achieved the highest ROC-AUC (0.955 ± 0.046), while Extra Trees obtained the highest Accuracy (0.878) and F1-score (0.887). Friedman analysis confirmed statistically significant differences among classifiers ($p<0.001$). Platt scaling consistently improved probabilistic reliability, whereas Beta calibration demonstrated stable performance under noise, feature perturbation, and reduced-data scenarios. Uncertainty-aware selective prediction increased high-confidence diagnostic accuracy by up to 8.1%, and decision curve analysis demonstrated improved clinical utility. The reliability analysis identified calibration-aware stacking as the most reliable ensemble configuration. An independent cross-dataset evaluation on a heterogeneous Alzheimer’s disease clinical dataset with a substantially different feature space yielded stable discrimination (ROC-AUC = 0.858 ± 0.025) and calibration (ECE = 0.132 ± 0.019) after the STACK_CAL architecture was independently retrained from scratch. These findings provide evidence of the cross-dataset applicability of the proposed reliability-aware strategy across different clinical data modalities, while further prospective and independent validation remains necessary before real-world clinical deployment. Full article

Driving style, reflected in recurrent patterns of acceleration, braking, speed selection, following distance, gear use, and lane-changing behavior, plays an important role in road safety and is also associated with fuel consumption, emissions, passenger comfort, and vehicle wear. This paper presents a structured review of recent research on driving style analysis, with particular emphasis on its relationship with road-safety outcomes and risk indicators. Following a PRISMA-oriented methodology, studies published between 2015 and 2025 were identified, screened, and synthesized to examine how driving styles are defined, detected, classified, and evaluated. The review shows a clear shift toward data-driven approaches, including feature-based machine learning and representation-learning methods using support vector machines, ensemble models, convolutional neural networks, recurrent neural networks, and hybrid deep learning architectures. Common data sources include smartphone inertial and GNSS signals, CAN/OBD vehicle data, telematics platforms, naturalistic driving datasets, and camera-based perception systems. Safety impact is most often assessed through crashes, near-miss events, traffic conflicts, time-to-collision measures, harsh maneuvers, and composite risk scores. Across the reviewed literature, aggressive and unstable driving patterns are generally associated with reduced safety margins and increased risk, although comparability remains limited by inconsistent label definitions, heterogeneous datasets, indirect safety proxies, and varied validation protocols. The paper also proposes a reproducible database architecture linking drivers, trips, driving events, and safety events to support transparent analysis, benchmark development, and future implementation in fleet monitoring, driver feedback, and connected vehicle applications. Full article

Background: Early detection of dementia is essential for timely intervention, yet existing diagnostic approaches remain costly, invasive, or dependent on specialized expertise. Electroencephalography (EEG) offers a non-invasive and accessible alternative; however, purely data-driven deep learning models may overlook clinically established neurophysiological biomarkers, particularly in the challenging detection of mild cognitive impairment (MCI). Methods: We propose the Clinical EEG Feature-Augmented Network (CEFA-Net), a knowledge-guided deep learning framework that systematically integrates automatic representation learning from raw multichannel EEG with clinically validated neurophysiological biomarkers. The architecture combines three complementary convolutional pathways capturing multi-scale temporal dynamics with domain-informed feature representations, enabling both data-driven discovery and clinically grounded interpretation. Task-specific optimization strategies—including focal loss, class-aware augmentation, and validation-guided ensemble weighting—were employed to enhance robustness under class imbalance. The model was evaluated on the large-scale the Chung-Ang University Hospital EEG (CAUEEG) dataset (1379 recordings from 1155 patients) across binary abnormality detection and three-class dementia staging tasks. Results: CEFA-Net achieved 81.02% accuracy (macro F1: 81.15%) for dementia staging and 87.15% accuracy (macro F1: 87.41%) for abnormality detection, outperforming baseline methods by 6.75–9.10 percentage points (p < 0.001). Notably, the proposed framework substantially improved MCI detection (F1-score: 78%), representing a 14-point gain over traditional machine learning approaches. Ablation analyses confirmed that clinical biomarker integration and multi-model fusion provide complementary diagnostic value. In an additional patient-disjoint evaluation using the no-overlap partitions, CEFA-Net achieved 85.40% accuracy for abnormality detection and 73.80% accuracy for dementia staging, demonstrating generalization to subjects completely excluded from the training data. Conclusions: These findings demonstrate that knowledge-guided integration of clinical biomarkers with deep representation learning can significantly enhance EEG-based dementia detection. CEFA-Net offers a clinically aligned and computationally efficient solution, supporting its potential for real-world screening and early diagnostic workflows. Full article

Construction material identification is important for automated construction monitoring, digital twin generation, and building information modelling. Terrestrial laser scanning (TLS) provides dense geometric information together with LiDAR intensity measurements; however, reliable material discrimination remains challenging because intensity is affected by acquisition geometry and surface texture may vary depending on the spatial scale at which it is characterised. This study proposes a TLS-only machine-learning framework combining a reflectance-related intensity–geometry regression descriptor with multiscale geometric roughness features. Plane-residual roughness and normal-variation roughness were calculated using local cube neighbourhoods with side lengths of 0.10, 0.20, and 0.30 m. To avoid ambiguity associated with surface-normal orientation, the normal-variation descriptor was calculated using orientation-invariant angular differences between neighbouring surface normals. The training dataset was balanced using distance-stratified random undersampling, while the held-out test dataset retained its original class distribution. FLAML was used for automated model selection and hyperparameter optimisation using three-fold cross-validation with macro F1-score as the optimisation metric. The final stacking classifier achieved an overall accuracy of 90.44%, balanced accuracy of 87.42%, and macro F1-score of 87.85% on 1,296,822 held-out test points. The held-out test data were acquired from different scanner positions and spatially distinct material regions from those used for training, with no shared point samples; however, both datasets originated from the same general study area, and broader cross-site generalisation therefore requires further independent validation. Most material classes showed strong discrimination, although Carpet remained challenging because of confusion with Asphalt. The results demonstrate the potential of combining TLS-derived radiometric information with multiscale geometric surface descriptors for construction material classification without relying on RGB colour information. Full article

With the escalating frequency of sophisticated zero-day attacks, overcoming the critical limitations of signature-based Intrusion Detection Systems (IDSs) has become paramount. This study proposes a hybrid multi-layered intrusion detection framework combining traditional machine learning, Deep Neural Architectures (DenseNN), and ensemble methods to evaluate zero-day resilience within cloud-level backend connectivity interfacing EV and V2X management ecosystems. Using the comprehensive CSE-CIC-IDS2018 benchmark as a surrogate environment, a code-executed Leave-One-Attack-Out (LOAO) cross-validation protocol across 13 distinct attack families was implemented to assess unseen-attack-family generalization within the benchmark to unseen threats. Furthermore, Explainable Artificial Intelligence (XAI) auditing, utilizing SHapley Additive exPlanations (SHAP) and Integrated Gradients, was integrated to inspect decision boundaries and resolve feature-attribution failure modes. Critically, the audit identified an artifact-driven data leakage caused by the Timestamp and identifier features, demonstrating that models learned temporal schedules rather than behavioral network signatures. Re-executing all experiments post-leakage removal quantified performance drops across all classifiers (e.g., Gaussian NB dropping by up to 20.88 percentage points in accuracy (at the 60% training ratio; 18.30 points at the 80% ratio)). Under standard binary classification metrics, tree ensembles (Random Forest and Extra Trees) achieved high in-distribution detection (F1 > 0.95) with rapid inference latency (≈0.05–−0.07 ms/sample). However, the rigorous LOAO evaluation revealed a substantial generalization penalty on truly unseen zero-day families (e.g., SQL Injection and Infiltration), where simpler linear models demonstrated broader generalization robustness (mean LOAO F1 = 0.397) compared with complex tree-ensemble models. By rectifying dataset leakage and benchmarking deployment trade-offs (training runtime, throughput, and memory footprint), this study delivers actionable, transparent guidelines for deployment-oriented IDS evaluation in dynamic network infrastructures. Full article

Background/Objectives: Polyendocrine metabolic ovarian syndrome (PMOS) is characterized by hyperandrogenism, particularly excessive testosterone, as a core clinical feature and a key pathogenic metabolite, yet its molecular mechanisms remain incompletely understood. Methods: This study integrated multi-omics data from Gene Expression Omnibus (GEO) databases with network toxicology, weighted gene co-expression network analysis (WGCNA), and machine learning to identify testosterone-associated core genes in PMOS. Results: Differential expression analysis and WGCNA yielded 42 candidate genes, from which five core genes, including GK5, CYP3A5, EGLN3, VCAM1, and AGTR1, were prioritized as top predictive features through ensemble modeling (RF + XGBoost). Molecular docking predicted favorable testosterone binding conformations. Regulatory network and drug enrichment analysis additionally predicted several upstream transcription factors, hub miRNAs, and potential repurposable drugs. Conclusions: These findings proposed a computational framework for a multi-target molecular landscape linking testosterone to PMOS. The identified genes, regulatory networks, and candidate drugs provided prioritized hypotheses for mechanistic exploration and future evaluation of potential diagnostic and therapeutic applications in hyperandrogenism-related PMOS. Full article

Reliable estimation of the compression index (Cc) is essential for settlement assessment, yet geotechnical databases often contain incomplete soil-index measurements. This study developed a missingness-aware heterogeneous ensemble that combined masked variables with binary availability indicators to predict Cc across eight predefined incomplete-input scenarios. The database comprised 1524 marine-clay specimens from eight coastal sites in South Korea. Five sites were used for model development and internal testing, while three sites were reserved for independent testing. A 12-dimensional representation allowed five artificial neural network seed models and four tree-based learners to process all scenarios using validation-derived weights. The proposed model achieved mean root mean square errors of 0.166 and 0.177 in the internal and independent tests, with corresponding coefficients of determination of 0.772 and 0.723. In the internal test, the proposed model produced more favorable point-estimate metric values than the case-specific artificial neural network and random forest baselines in all 32 comparisons and than XGBoost in 30 comparisons. The corresponding differences were less consistent in the independent test, and only six of the 24 unadjusted bootstrap confidence intervals remained entirely below zero. The framework provides a unified tool for preliminary Cc screening under the predefined incomplete-input scenarios evaluated in this study. Full article

Atmospheric river (AR)-driven flooding is a natural hazard that causes severe damage in many regions, and the damage escalates sharply once an AR persists beyond a certain duration. Existing studies and numerical weather prediction models, however, have focused mainly on AR occurrence and on the intensity of integrated vapor transport (IVT) at individual time steps, paying little attention to duration. California and Chile are both exposed to severe AR-related hazards, yet for the period since 2000, the Global Ensemble Forecast System (GEFS) forecast at a two-day lead, compared against the regional IVT threshold, correctly predicts persistence for only about 30% of the ARs that actually persisted for 24 h or longer. This study retains the predictive capability of the physics-based GEFS forecast while correcting its weak performance on persistence using TabPFN, a pre-trained transformer for tabular data. Using single-control-member forecasts from the GEFS v12 reforecast (2000 to 2019), the model predicts the minimum IVT over the target window and compares it with the regional threshold. In the performance evaluation, the F1 score, which combines the precision and recall of AR persistence prediction into a single measure, rose from 0.414 to 0.502 and 0.601 in the two regions, and TabPFN outperformed machine learning models such as 1D-CNN and LGBM. Moreover, on the California coast, the proposed model, through its persistence decisions, captured 66.3% of the rainfall that fell during persistent atmospheric river events, up from 30.7% for the raw forecast. The proposed model offers a forecast post-processing method for predicting AR persistence and can contribute meaningfully to flood disaster prevention. Full article

As bandwidth-intensive applications proliferate and the usage of wireless devices surges, fifth-generation (5G) and beyond (B5G) networks are challenged to enhance coverage, reduce latency, and improve efficiency. The application of machine learning (ML) models for received signal strength (RSS) estimation is a powerful tool. This study evaluates various ML models—categorical boosting (CatBoost), extreme randomized trees (ETs), light gradient boosting machine (LGBM), and extreme gradient boosting (XGBoost)—for effective estimation of RSS. Additionally, we apply explainable artificial intelligence (XAI) methodologies, especially the Shapley additive explanation (SHAP) framework. Our investigation reveals the sophisticated mechanisms within these models, notably highlighting the exceptional accuracy of the ET model. We further introduce SHAPRP, in which SHAP attributions reduce the input space and sparse regression selects a compact subset of the ET ensemble. The results are obtained from a single-operator rural/semi-rural campaign and constitute a case study, where the trained estimator is deployment-specific and is not a pre-trained model applicable to arbitrary 5G/B5G scenarios, so what transfers is SHAPRP itself. Full article

Neonatal sepsis remains a major cause of morbidity and mortality worldwide, while timely diagnosis continues to be challenging because of nonspecific clinical manifestations and limitations of conventional diagnostic methods. Recent advances in artificial intelligence (AI) have created new opportunities for the early prediction and diagnosis of neonatal sepsis through the analysis of large and complex clinical datasets. This structured narrative review summarizes current evidence regarding AI-based approaches for neonatal sepsis prediction and diagnosis. A literature search of PubMed, Scopus, and Google Scholar identified studies evaluating machine learning, deep learning, and advanced predictive analytics using clinical, laboratory, physiological, electronic health record, and multi-omics data. Current evidence suggests that AI models, particularly ensemble learning, gradient boosting, and deep learning approaches, can achieve promising predictive performance and identify infants at increased risk of sepsis hours before conventional clinical recognition. Continuous physiological monitoring and multimodal data integration appear particularly promising for real-time prediction. However, important challenges remain, including limited external validation, small and heterogeneous datasets, concerns regarding interpretability, and unresolved ethical and regulatory issues. Future progress will depend on multicenter collaboration, explainable AI frameworks, federated learning, and multimodal predictive models. Although current evidence supports the predictive potential of AI-based models, prospective multicenter validation and clinical impact studies are required before improvements in neonatal clinical outcomes can be established. Artificial intelligence has the potential to become a valuable clinical decision support tool to support early sepsis recognition and precision neonatal care. Full article

Background: Port access traffic combines pronounced calendar structure with substantial day-to-day variability, raising the question of whether temporal heterogeneity should condition a common forecasting configuration or support model specialization. Methods: Daily car and freight-truck arrivals at an urban non-containerized port were forecast at 1-, 7-, and 14-day horizons using statistical, machine-learning, and composite models under a leakage-free rolling-origin protocol. A unified calendar-conditioned ensemble was compared with a prespecified regime-based model assignment, with development, test, and temporal stress-test periods kept distinct. Results: The unified ensemble achieved next-day R² values of 0.75 for trucks and 0.80 for cars on the held-out 2018 test period and 0.73 and 0.75, respectively, under the 2019 temporal stress-test. Model preference varied jointly with temporal regime, forecast horizon, and evaluation period. Although fixed assignment improved car accuracy on the 2018 test period, the model–regime dominance required to justify it was not sufficiently stable across horizons and evaluation periods. Conclusions: Calendar conditioning provides a robust strategy for the present forecasting problem. More generally, fixed regime-based specialization should be supported by sufficiently stable model–regime dominance rather than inferred from identifiable temporal heterogeneity alone. Full article

Background/Objectives: Healthcare service quality is a critical dimension of patient safety and operational performance. Traditional assessment approaches are often limited in their ability to support prediction, which has increased interest in machine learning (ML) models. This systematic review assessed the effectiveness of ML algorithms in predicting healthcare service quality metrics, with emphasis on their applications, comparative performance and implementation challenges. Methods: A systematic literature search was conducted in PubMed/MEDLINE, Scopus, CINAHL, and ScienceDirect for studies published between 1 January 2020 and 30 July 2025. Following title/abstract screening and full-text review, 49 studies were included. Studies were grouped into conventional/ensemble ML and deep learning categories based on the primary model class analyzed in each article. Results: The reviewed studies focused mainly on acute clinical and operational outcomes, especially length of stay (27.0%), mortality (23.0%), and readmission rates (18.0%), while subjective, patient-centered metrics received less attention. Conventional and ensemble ML models, particularly RF and XGBoost, were frequently reported, while deep learning models were used in more complex prediction tasks. Conclusions: The evidence suggests that well-validated and interpretable ML models can support healthcare quality prediction. However, important challenges remain regarding implementation, validation, generalizability, and data heterogeneity. Full article

Accurate wind power forecasting is essential for the stable and economic operation of power systems with high renewable penetration. Although machine learning models have been widely adopted for this task, the assumption that greater model complexity invariably yields superior forecasting accuracy has received insufficient scrutiny. This paper presents a systematic experimental investigation that covers two complementary stages, i.e., wind speed correction and wind power forecasting. For wind speed correction, we compare 10 machine learning methods, including spanning linear, instance-based, and tree-based ensemble learners, under four newly proposed progressively enriched feature configurations. For wind power forecasting, we benchmark 20 methods spanning traditional machine learning, time-series deep learning, and Transformer-based architectures on two geographically distinct wind farms. Our results reveal a clear task-dependent pattern. In wind speed correction, tree-based ensemble methods, particularly gradient boosting variants, consistently dominate, and feature engineering contributes more to accuracy gains than model selection. In wind power forecasting, deep learning architectures substantially and consistently outperform traditional methods, with attention-based models generalizing the most robustly across regimes and recurrent networks proving to be the most sensitive to regime shifts. These findings provide actionable task-specific guidance for model selection in operational wind power forecasting systems. Full article