Search Results (5,497)

Leakage identification in district heating networks is challenging because leakage-induced transient pressure waves often overlap with pressure disturbances triggered by routine operations such as valve regulation, pump speed variation, and emergency shut-off. In addition, the scarcity of high-quality labeled leakage samples limits the robustness of data-driven models under small-sample conditions. To address these issues, this study proposes a compact and moderately interpretable framework for multi-condition identification from transient pressure-wave signals, integrating signal preprocessing, handcrafted statistical feature extraction, multiclass ReliefF-based feature selection, and class-wise generative adversarial network augmentation in the selected feature space. A dataset containing four representative conditions, namely leakage, valve regulation, pump speed regulation, and emergency valve shut-off, was constructed using an integrated indoor district heating network testbed. After Hampel-based spike suppression and zero-phase Butterworth band-pass filtering within 0.5 to 300 Hz, time- and frequency-domain statistical features were extracted, and a compact subset was selected by multiclass ReliefF. A class-wise generative adversarial network was then used to augment the training set in feature space, while all evaluations were performed strictly on real samples. The results show that feature-space augmentation improves robustness and generalization under operational disturbances and noise. Using random forest as the representative classifier, Accuracy and Macro-F1 increased from 0.960 to 0.985, while leakage recall improved from 0.920 to 0.980. Further comparisons confirmed that the ReliefF-selected subset outperformed representative alternatives such as LASSO and mRMR. Overall, the proposed framework provides an effective solution for distinguishing leakage events from operational disturbances and offers practical support for online monitoring and intelligent operation of district heating networks. Full article

Background: During the flu season, there is an increase in absenteeism due to illness, a drop in productivity, and a greater risk of the virus spreading among workers. Thus, the Italian Ministry of Health recommends vaccination for essential service workers. The University of Salerno, in collaboration with the local health authority of Salerno, offers free vaccination to its employees. Methods: A public health methodology for seasonal influenza vaccination in the workplace is presented—specifically in the university setting—with the aim of identifying individual, contextual, and organizational elements of the model that have promoted vaccination uptake. An ad hoc questionnaire was used (October–December 2025) to survey 399 academic employees, investigating seasonal influenza vaccination in the following aspects: recent personal experiences, motivations, vaccination experiences at university, sources of information, considerations regarding national and local vaccination campaigns, and level of vaccine confidence (VCI). Results: Seasonal influenza vaccination at the University is appreciated for its compatibility with working hours (66.1%), the availability of a platform that allows flexible booking (56.9%), the perception of safety in the environment (31.6%), the fact that the vaccine is free (17.4%), and the involvement of office/laboratory colleagues (5%). Participants appreciate the model and would apply it to other vaccinations at the University and in other institutional settings. A significant relationship (F = 7.24; df = 1; p < 0.05) exists between confidence in the vaccine and the sense of security experienced when receiving the vaccine in the workplace. Data analysis was performed using the IBM SPSS v.28 software. Conclusions: The model proposed can be applied to other institutional contexts, simplifying and facilitating access to vaccines by implementing vaccination campaigns tailored to specific work environments. Full article

Background: Lung cancer remains one of the leading causes of cancer-related mortality worldwide, primarily due to late diagnosis. Although machine learning (ML) techniques have been widely applied for lung cancer classification, many studies lack a fully optimized end-to-end pipeline using routine clinical data. This study proposes an optimized ML framework that integrates demographic, lifestyle, and clinical features with systematic hyperparameter tuning to improve classification performance. Methods: A dataset of 309 patient records containing demographic, lifestyle, and clinical attributes was used. The data were preprocessed and split into training and testing sets in an 80:20 ratio. Feature selection was performed using metaheuristic algorithms, including Red Deer Optimization, Binary Grasshopper Optimization, Gray Wolf Optimization, and Bee Colony Optimization. Six ML classifiers—Logistic Regression, Support Vector Classifier, Gradient Boosting, Random Forest, K-Nearest Neighbors, and Gaussian Naive Bayes—were trained with optimized hyperparameters. Model performance was evaluated using accuracy, precision, recall, F1-score, and ROC–AUC. Results: The optimized pipeline significantly improved classification performance. Logistic Regression achieved the highest accuracy of 91.07% with an AUC of 0.91, outperforming more complex ensemble models. Gradient Boosting and Random Forest both achieved an accuracy of 87.5%, while other classifiers demonstrated moderate performance. Conclusions: The proposed optimized ML pipeline enhances lung cancer classification accuracy using routine clinical data. The results highlight that simpler, well-optimized models can outperform complex approaches on structured datasets. This framework shows strong potential for early lung cancer risk screening and clinical decision support, although further validation on larger datasets is recommended. Full article

Search and rescue (SAR) missions in collapsed or underground environments remain challenging due to GPS unavailability, which hinders localization and autonomous navigation. Systems that rely on single-sensor inputs or structured settings often degrade under smoke, dust, or dynamic clutter. This paper presents an autonomous ground robot for GPS-denied SAR that integrates low-cost thermal, visual, inertial, and acoustic cues within a unified, computation-efficient architecture. The stack combines Thermal–Visual Odometry (TV–VO) with Zero-Velocity Updates (ZUPT) for drift-resistant localization, RescueGraph for multimodal survivor detection, and a Proximal Policy Optimization (PPO) planner for adaptive navigation under uncertainty. Across simulated disaster scenarios and benchmark corridor runs, the system shows embedded-feasible runtime behavior and supports return to base without external beacons under the evaluated conditions. Quantitatively, TV–VO+ZUPT reduces drift in short internal evaluations, while RescueGraph attains an F1-score of 0.6923 and an area under the ROC curve (AUC) of 0.976 for survivor detection. At the system level, the integrated navigation stack achieves full mission completion in the reported SAR-style trials, while the separate A*/PPO comparison highlights a trade-off between completion rate, traversal time, and collisions. Overall, the results support the practical promise of a low-cost sensor-fusion and learning-assisted navigation framework for GPS-denied SAR robotics. Full article

Legal element recognition, which identifies discrete factual elements in Chinese court judgments to support judicial analysis and case retrieval, faces a severe long-tail challenge: head-to-tail label-frequency ratios exceed 100:1, and over 60% of sentences carry no label, starving rare elements of training signal. Static reweighting methods assign fixed weights prior to training and cannot respond to the model’s evolving confidence; sample-level meta-learning couples all co-occurring label gradients to a single scalar, preventing independent tail-label amplification. We propose BML-Trans, a boundary-aware meta-learning framework that addresses both limitations. A label-wise meta-weighting mechanism maintains per-label gradient weights updated via bilevel hypergradient descent, decoupling tail-label amplification from co-occurring head labels. A boundary-aware meta-set concentrates calibration signal on high-uncertainty, tail-triggering sentences rather than on easy negatives, and a lightweight Multi-Scale Adapter sharpens the warm-up probability estimates on which boundary selection depends. Concretely, BML-Trans achieves an average Avg-F1 of 82.5% on CAIL2019 across the labor, divorce, and loan domains, outperforming the strongest baseline by 1.2 percentage points overall and by up to 5.7 percentage points on tail-label Macro-F1, at only 14% additional training cost. Ablation confirms a cascade dependency among the three components, establishing that the gains are structural rather than incidental to threshold selection or initialization. Full article

Background/Objectives: We developed and validated a multimodal magnetic resonance imaging (MRI) framework combining deep learning segmentation with radiomics to predict local recurrence in nasopharyngeal carcinoma (NPC). Methods: This retrospective two-center study included 1074 NPC patients treated between 2015 and 2019. Center 1 cases were split 8:2 into training and internal test sets, while Center 2 served for external validation. A multimodal Swin UNet model automatically segmented tumors from pretreatment T1-weighted, T2-weighted, and contrast-enhanced T1 (CET1) images. Radiomics features were extracted from expert-reviewed regions of interest, selected, and modeled using extreme gradient boosting for recurrence prediction. Results: The multimodal segmentation model maintained consistent but moderate Dice similarity coefficients (0.737, 0.666, and 0.726 for T1WI, T2WI, and CET1 in external validation). These values reflect the moderate overlap typical for nasopharyngeal carcinoma, given its highly infiltrative growth and ill-defined boundaries along complex anatomic interfaces. For local recurrence prediction, single-modality models reached external AUCs between 0.754 and 0.781. Importantly, the multimodal fusion model demonstrated numerical improvement over single modalities in the external validation set (e.g., vs. T1WI, p = 0.141), achieving an AUC of 0.910, accuracy of 0.908, sensitivity of 0.805, specificity of 0.946, and F1-score of 0.825. Conclusions: The multimodal MRI radiomics model, developed alongside a deep learning segmentation module, demonstrated favorable multicenter performance for evaluating NPC recurrence risk. The primary prognostic analysis was based on expert-reviewed regions of interest; a supplementary analysis using fully automatic segmentation masks yielded comparable, non-significantly different performance across all cohorts (Training AUC: 0.887; Internal Test AUC: 0.892; External Validation AUC: 0.885 vs. 0.910, p = 0.145), supporting the feasibility of future end-to-end deployment. Fusing multimodal features yielded numerical improvements over single-sequence models in external validation, providing a basis for post-treatment surveillance planning. Full article

Background: Cysteine proteases of Clonorchis sinensis are potential diagnostic antigens, yet the performance of individual members within this diverse enzyme family requires systematic evaluation. This study aimed to assess the diagnostic potential of four recombinant cysteine proteases (rCsCP1–4) for human clonorchiasis. Methods: An indirect ELISA was developed to measure serum reactivity (IgG, IgG subclasses, IgA) against rCsCP1–4. The assay was validated using 180 microscopy-confirmed positive and 148 negative control sera. Samples were randomly split into training and validation sets (7.5:2.5). Diagnostic performance of single antigens and their combinations was evaluated using univariate and multivariate logistic regression and compared with a commercial kit. Key metrics included the area under the curve (AUC), sensitivity, specificity, accuracy, F1-score, and Kappa coefficient. Results: Four single antigen–antibody pairs showed high performance: rCsCP1-IgG4 (AUC = 0.928), rCsCP2-IgA (AUC = 0.863), rCsCP3-IgG1 (AUC = 0.920), and rCsCP4-IgG4 (AUC = 0.958). Among these, rCsCP1-IgG4, rCsCP3-IgG1, and rCsCP4-IgG4 outperformed the commercial kit, achieving higher sensitivity (92.0%, 96.0%, 96.0% vs. 86.0%), specificity (87.5%, 81.3%, 90.6% vs. 78.1%), accuracy (92.0%, 88.9%, 94.1% vs. 86.0%), and F1-scores (0.902, 0.902, 0.939 vs. 0.829). The Kappa values for rCsCP1-IgG4 (0.768) and rCsCP4-IgG4 (0.773) indicated substantial agreement with the microscopic standard. Multi-antigen combinations (triple or quadruple) further enhanced performance, achieving sensitivity and specificity > 98% with an AUC approaching 1.0. Conclusions: This study identifies rCsCP1 and rCsCP4, particularly in combination with IgG4 detection, as highly promising diagnostic targets for clonorchiasis. Multi-antigen combinations significantly improved diagnostic performance compared to single-antigen assays, offering a strategy for high-precision diagnosis. Furthermore, the efficacy of the rCsCP2-IgA pair suggests that detecting fecal secretory IgA could be a novel avenue for non-invasive, self-testing applications. Full article

The proliferation of fake news across social media, headlines, and news articles poses major challenges for automated detection, particularly in multilingual and cross-media settings affected by data imbalance. We propose a fake news detection framework based on LLM-driven, feature-guided text augmentation. The method generates realistic synthetic samples across languages, media types, and text granularities while preserving meaning and stylistic coherence. Experiments with classical and transformer-based models (Random Forest, Logistic Regression, BERT, XLM-R) across social media, headlines, and multilingual news datasets show consistent improvements in performance. For inherently balanced datasets (e.g., social media), synthetic augmentation yields negligible but stable performance changes. Across imbalanced scenarios, synthetic augmentation substantially improves minority-class recall and F1-score (e.g., fake news recall from 0.57 to 0.86), while preserving majority-class performance, leading to more balanced and reliable classifiers, whereas oversampling significantly degrades results due to overfitting on duplicated language patterns. Overall, a hybrid semantic- and style-based model proves to be the most robust strategy, outperforming oversampling and matching or exceeding baseline performance across datasets. Full article

This article investigates hop and 2-hop domination numbers, each defined by the minimum of its hop (2-hop) dominating sets on the honeycomb and butterfly networks, which have enormous applications in chemistry, engineering, parallel computing, etc. Furthermore, we obtain an upper bound on a regular honeycomb mesh network $HC(n,k)$ and a lower bound for an n-dimensional butterfly network $BF\left(n\right)$ with $(n+1)$ levels/stages. The cases in which these determined bounds are tight are also explored. Full article

The purpose of this research work is to propose and develop the notion of $\left(\alpha ,\psi \right)$-contractions in the setting of extended $\mathcal{F}$-metric spaces and to establish corresponding fixed point results. Using these results, we derive fixed point results for graphic contractions in extended $\mathcal{F}$-metric spaces as well as for mappings in partially ordered extended $\mathcal{F}$-metric spaces. To demonstrate the validity and novelty of the proposed results, a non-trivial example is provided. Moreover, the constructed framework serves as a tool to investigate the existence of solutions for Caputo fractional differential equations, thereby highlighting both its effectiveness and practical significance. Full article

Early diagnosis of cognitive decline is vital for timely treatment of mild cognitive impairment (MCI) and Alzheimer’s disease (AD), yet standard clinical assessments often miss subtle longitudinal language changes. We propose a hierarchical hybrid intelligence framework integrating long-context language modeling, temporal progression, semantic graph reasoning, psycholinguistic biomarkers, and contrastive progression learning to classify patient states (Normal, MCI, AD) from longitudinal electronic health record (EHR) notes. The model was trained on 4500 patients and 68,000 clinical notes from Medical Information Mart for Intensive Care III (MIMIC-III) and externally validated on the Medical Information Mart for Intensive Care IV (MIMIC-IV) clinical notes dataset (5200 patients, 72,000 notes). Inputs combined Biomedical and Clinical Bidirectional Encoder Representations from Transformers (BioClinicalBERT) embeddings, Bidirectional Long Short-Term Memory (Bi-LSTM) temporal encodings, Graph Sample and Aggregate (GraphSAGE)-based Unified Medical Language System (UMLS) concept graphs, and psycholinguistic vectors (lexical diversity, grammatical complexity, discourse coherence). On the MIMIC-III hold-out set, the model achieved 99.999% accuracy, a macro F1-score of 0.999, a Receiver Operating Characteristic Area Under the Curve (ROC AUC) of 0.999, and a temporal stability variance of 0.0008. Monte Carlo cross-validation (10,000 folds) yielded $99.997\pm 0.003$% accuracy and $0.999\pm 0.001$ macro F1. Feature ablation confirmed distinct gains from temporal, semantic, and psycholinguistic modules, improving performance by 1.1% over text-only baselines. Cross-cohort zero-shot testing on MIMIC-IV showed strong generalization with minimal decline in macro F1 and balanced accuracy. Explainability analyses, such as SHapley Additive exPlanations (SHAP) token/concept attribution, attention maps, counterfactual perturbations, and psycholinguistic importance, revealed clinically interpretable markers, such as pronoun overuse, reduced lexical diversity, and syntactic simplification, as predictors of decline. Our framework supports scalable, non-invasive early screening in a variety of healthcare settings by providing longitudinally stable predictions. Full article

Introduction: Hospital adverse events are a global patient safety problem that account for avoidable death, long-term disability, extended length of stay, and increased healthcare costs. Underreporting, wherein fewer than 10% of events are indeed recorded, is prevalent and is characterized primarily by cultural and organizational determinants. Artificial intelligence, in the form of machine learning and natural language processing, has been described as a potential tool for enhancing adverse events detection and prediction with the use of large-scale clinical data. Materials and Methods: PRISMA-DTA guidelines were followed in this systematic review. Scopus, PubMed, and Web of Science were searched employing keywords associated with adverse events, artificial intelligence methodologies (e.g., machine learning, deep learning, natural language processing), and healthcare settings. Inclusion criteria included original research on artificial intelligence-based solutions for the detection or prediction of adverse events such as medication errors, hospital-acquired infections, and complications during surgery. Reviews, meta-analyses, and non-artificial intelligence studies were excluded. Following screening, 15 studies were found to meet inclusion criteria. Results: The referenced studies show a shift from rule-based natural language processing models to advanced deep learning and Bidirectional Encoder Representations from Transformers models. Early approaches, i.e., Support Vector Machine classifiers, achieved AUC scores as high as 0.92, while later models (Random Forest, LightGBM, XGBoost) mirrored AUCs of over 0.93. Large language models achieved F1-scores of 0.84 for named entity recognition. Artificial intelligence models even identified unreported incidents. Discussion: Artificial intelligence-powered methods are transforming adverse events detection from retrospective to predictive, proactive monitoring. There remain some challenges, however, including limited external validation, class imbalance, and interpretability of complex models. Future studies must address explainable artificial intelligence, multicenter trials, and high-quality well-annotated datasets to offer secure clinical integration. Full article

Multiple sclerosis (MS) is an inflammatory and neurodegenerative disease. Optical coherence tomography (OCT) is a non-invasive imaging technique of the retina. The thickness of the ganglion cell–inner plexiform layer (GCIPL) obtained from an OCT image is a valuable biomarker for monitoring MS. Since the functional shape (F-shape)-based technique has proven to be an effective platform for detecting glaucoma using OCT images, in this paper, we develop an F-shape-based framework to distinguish MS subjects from healthy ones using the thickness of GCIPL. The thickness of the GCIPL layers in the macula region of OCT images in a selected region of interest (ROI) for a set of healthy and MS subjects is represented as F-shape objects, which are registered to a common template using atlas registration. The residual F-shapes, defined as the difference between the F-shape of this common template and the individual registered F-shapes, are used to train an support vector machine (SVM) classifier and subsequently to detect MS. Accuracy, sensitivity, specificity, and area under the curve (AUC) are used to evaluate and compare the classification performance of the proposed F-shape-based scheme and those of sectoral-based schemes. The proposed F-shape-based scheme is shown to significantly outperform the sectoral-based schemes. The superior performance of the proposed F-shape-based scheme can be attributed to the use of (i) a highly dense mesh formed on the ROI in the macula region, (ii) atlas registration that puts the F-shapes of all the subjects on a common platform, and (iii) residual thicknesses as input features for the classification. Full article

Background: Metabolic syndrome (MetS) leads to alterations in cardiac autonomic control that can be detected from electrocardiogram (ECG)-derived markers, particularly when the cardiovascular system is challenged during an oral glucose tolerance test (OGTT). Methods: In this paper, we present an automated framework for MetS identification using RR intervals and heart rate variability (HRV) features extracted from 12-lead ECG recordings acquired during the five OGTT stages in 40 male participants (15 with MetS, 10 controls, and 15 endurance-trained marathon runners). RR intervals were first derived using a multilead Pan-Tompkins approach with fusion-based validation. From these RR series, HRV descriptors were computed from time-domain statistics (RR mean, SDNN, rMSSD, pNN50), spectral indices (VLF, LF, HF, LF/HF), and nonlinear measures (SD1, SD2, SampEn, DFA-${\alpha }_1$). Conventional HRV analysis revealed pronounced physiological differences between groups: MetS subjects exhibited reduced parasympathetic activity, reflected by lower rMSSD and SD1, lower HF power, and higher LF/HF ratios, whereas marathoners showed greater vagal modulation, higher HF power, and increased signal complexity. Healthy controls showed an intermediate autonomic profile. Using RR sequences and HRV descriptors (256 samples per stage), we trained three multimodal classifiers: a CNN-MLP model with a softmax output, a CNN-MLP model with an SVM head, and a CNN + LSTM-MLP + SVM architecture. Results: All models achieved strong discriminative performance, with accuracies ranging from 0.92 to 0.95, F1-macro values from 0.92 to 0.95, and macro-AUC values from 0.96 to 0.97. The CNN-MLP model achieved the best overall performance, whereas the CNN + LSTM-MLP + SVM model showed strong class discrimination, particularly for endurance athletes, while maintaining competitive recall for MetS. Conclusions: These findings support the feasibility of ECG-based autonomic assessment as a complementary non-invasive approach for early metabolic risk detection in clinical and preventive cardiometabolic screening settings. Full article

Breast cancer histopathology classification remains a fundamental challenge in computational pathology due to variations in tissue morphology across magnification levels. Convolutional neural networks (CNNs) have long been the standard for image-based diagnosis, yet recent advances in vision-language models (VLMs) suggest they may provide strong and transferable representations for complex medical images. In this study, we present a systematic comparison between CNN baselines and large VLMs—Qwen2 and SmolVLM—fine-tuned with Low-Rank Adaptation (LoRA; $r=16$, $\alpha =32$, dropout = $0.05$) on the BreakHis dataset. Models were evaluated at $40\times$, $100\times$, $200\times$, and $400\times$ magnifications using accuracy, precision, recall, F1-score, and area under the ROC curve (AUC). While Qwen2 achieved moderate performance across magnifications (e.g., 0.8736 accuracy and 0.9552 AUC at $200\times$), SmolVLM consistently outperformed Qwen2 and substantially reduced the gap with CNN baselines, reaching up to 0.9453 accuracy and 0.9572 F1-score at $200\times$—approaching the performance of AlexNet (0.9543 accuracy) at the same magnification. CNN baselines, particularly ResNet34, remained the strongest models overall, achieving the highest performance across all magnifications (e.g., 0.9879 accuracy and 0.9984 AUC at $40\times$). These findings demonstrate that LoRA fine-tuned VLMs, despite requiring gradient accumulation and memory-efficient optimizers and operating with a significantly smaller number of trainable parameters, can achieve competitive performance relative to traditional CNNs. However, CNN-based architectures still provide the highest accuracy and robustness for histopathology classification. Our results highlight the potential of VLMs as parameter-efficient alternatives for digital pathology tasks, particularly in resource-constrained settings. Full article