Search Results (80)

Parkinson’s Disease (PD) is a progressive neurodegenerative disorder for which early and reliable diagnosis remains challenging. To address this challenge, the key innovation of this work is a confidence-gated fusion mechanism that dynamically weights classifier outputs based on per-sample prediction certainty, overcoming the limitations of static ensemble strategies. Building on this idea, we propose a Confidence-Gated Hybrid CNN Ensemble that integrates CNN-based acoustic feature extraction with heterogeneous classifiers, including XGBoost, Support Vector Machines, and Random Forest. By adaptively modulating the contribution of each classifier at the sample level, the proposed framework enhances robustness against data imbalance, inter-speaker variability, and feature complexity. The method is evaluated on two benchmark PD speech datasets, where it consistently outperforms conventional machine learning and ensemble approaches, achieving a best classification accuracy of up to 97.9% while maintaining computational efficiency compatible with real-time deployment. These results highlight the effectiveness and clinical potential of confidence-aware ensemble learning for non-invasive PD detection. Full article

Entity alignment (EA) facilitates knowledge fusion by matching semantically identical entities in distinct knowledge graphs (KGs). Existing embedding-based methods rely solely on intrinsic KG facts and often struggle with long-tail entities due to insufficient information. Recently, large language models (LLMs), empowered by rich background knowledge and strong reasoning abilities, have shown promise for EA. However, most current LLM-enhanced approaches follow the in-context learning paradigm, requiring multi-round interactions with carefully designed prompts to perform additional auxiliary operations, which leads to substantial computational overhead. Moreover, they fail to fully exploit the complementary strengths of embedding-based small models and LLMs. To address these limitations, we propose HIEA, a novel hierarchical inference framework for entity alignment. By instruction-tuning a generative LLM with a unified and concise prompt and a knowledge adapter, HIEA produces alignment results with a single LLM invocation. Meanwhile, embedding-based small models not only generate candidate entities but also support the LLM through data augmentation and certainty-aware source entity classification, fostering deeper collaboration between small models and LLMs. Extensive experiments on both standard and highly heterogeneous benchmarks demonstrate that HIEA consistently outperforms existing embedding-based and LLM-enhanced methods, achieving absolute Hits@1 improvements of up to 5.6%, while significantly reducing inference cost. Full article

Money laundering poses a serious threat to financial stability and requires effective national frameworks for prevention. This study investigates how the quality of legal and institutional frameworks affects the effectiveness of national anti-money laundering (AML) systems and their implications for financial risk management. We conducted an empirical analysis of 132 jurisdictions in 2024 using the Basel AML Index (AMLI) and the WJP Rule of Law Index (RLI). The Random Forest method was employed to model the relationship between rule-of-law indicators and AML risk levels. Findings reveal a significant inverse relationship between rule-of-law indicators and AML risk levels, with an overall classification accuracy of 69.6%. The model performed best for low-risk countries (precision 75%, recall 92.31%), moderately for medium-risk countries (precision 65.22%, recall 78.95%), but failed to identify high-risk jurisdictions, suggesting a legal institutional “threshold” necessary for effective AML functioning. Key predictors included protection of fundamental rights and mechanisms for civil oversight, with strong negative correlations between AML risk and criminal justice impartiality (−0.35), civil justice fairness (−0.35), and equality before the law (−0.41). These results show that legal factors strongly affect AML risk and can guide regulators in improving risk-based standards, enhancing regulatory certainty, and managing financial risk. Full article

This article investigates the probabilistic relationship between quantum classification of Boolean functions and their Hamming distance. By integrating concepts from quantum computing, information theory, and combinatorics, we explore how Hamming distance serves as a metric for analyzing deviations in function classification. Our extensive experimental results confirm that the Hamming distance is a pivotal metric for validating nearest neighbors in the process of classifying random functions. One of the significant conclusions we arrived is that the successful classification probability decreases monotonically with the Hamming distance. However, key exceptions were found in specific classes, revealing intra-class heterogeneity. We have established that these deviations are not random but are systemic and predictable. Furthermore, we were able to quantify these irregularities, turning potential errors into manageable phenomena. The most important novelty of this work is the demarcation, for the first time to the best of our knowledge, of precise Hamming distance intervals for the classification probability. These intervals bound the possible values the probability can assume, and provide a new foundational tool for probabilistic assessment in quantum classification. Practitioners can now endorse classification results with high certainty or dismiss them with confidence. This framework can significantly enhance any quantum classification algorithm’s reliability and decision-making capability. Full article

Background: Footwear may influence paediatric gait biomechanics, yet evidence across footwear types and barefoot conditions remains heterogeneous. This review aimed to synthesise evidence on how footwear affects gait biomechanics in children and adolescents compared with barefoot walking, and to conduct exploratory meta-analyses when feasible. Methods: We performed a PRISMA-guided systematic review (PubMed and Scopus; inception to November 2025) including participants < 18 years with gait outcomes assessed under barefoot and/or defined footwear conditions. Outcomes included spatiotemporal, kinematic, kinetic, and plantar-pressure variables. Risk of bias was assessed with ROBINS-I. Random-effects meta-analyses (inverse-variance) were conducted only when ≥2 studies reported comparable outcomes. Results: Twenty-two studies were included; most were observational with overall moderate-to-serious risk of bias, mainly due to confounding and participant selection. Quantitative synthesis (exploratory): Meta-analyses were possible only for ankle plantarflexion (k = 2) and stride length (k = 3) and showed non-significant pooled effects with extreme heterogeneity (I² > 90%) and wide prediction intervals. Narrative synthesis: For other outcomes, heterogeneity in designs, footwear definitions, and measurement protocols precluded pooling; conventional footwear may reduce metatarsophalangeal mobility and alter kinematics and plantar pressure in some contexts, while minimalist/biomimetic features may approximate barefoot-like values for selected parameters without implying equivalence. Conclusions: Footwear exposure in childhood may affect several gait-related parameters, but the certainty of evidence is low to moderate due to risk of bias, inconsistency, and imprecision. Standardised footwear classifications, harmonised gait protocols, and longitudinal studies are needed to clarify developmental implications and inform evidence-based guidance. Full article

Forests are a critical component of global carbon sequestration, biodiversity, and ecosystem services, making accurate mapping essential for long-term monitoring. In North Korea, limited field access, rugged topography, and inconsistent national statistics necessitate reliable remote sensing–based observation. However, frequent cloud contamination challenges the use of optical time-series imagery for forest monitoring. This study introduces a framework that integrates a ConvLSTM-based autoencoder into a U-Net segmentation model to improve tree cover classification from Sentinel-2 time-series data. The autoencoder was pretrained to reconstruct cloud-contaminated or missing observations using multi-octave Perlin-noise perturbations, providing standardized inputs that enhanced segmentation robustness under noisy conditions. Results show that tree cover accuracy exceeded 96% when all five time steps were available and remained stable (94–95%) even with one missing step. Accuracy declined below 90% with three missing steps but remained above 80%, enabling draft classifications under limited data. Confidence analysis further indicated that model certainty is a practical quality-control metric. Annual mapping for 2019–2024 showed a general increase in tree cover, aligning with reported afforestation efforts in North Korea. Taken together, the framework advances long-term monitoring, carbon accounting, and risk assessment in North Korea, while also enabling robust, region-adapted monitoring in cloud-prone, data-limited settings. Full article

Precision medicine starts with a precision diagnosis. Yet up to 80% of cases of monogenic diabetes, a form of diabetes characterized by mutations in a single gene, are either overlooked or misdiagnosed. A genetic test for monogenic diabetes does not always lead to a precise diagnosis, as novel variants are often classified as variants of unknown significance. Variant interpretation requires collation of a framework of evidence, including population, computational, and segregation data, and can be assisted by functional analysis. The inclusion of functional data can be challenging, depending on the number of benign and pathogenic variants available for benchmarking assays. Glucokinase is the rate-limiting step for glucose metabolism in the pancreatic beta-cell and governs the threshold for glucose-stimulated insulin release. Loss-of-function alleles in the glucokinase (GCK) gene are a cause of stable fasting hyperglycemia from birth and/or diabetes. In this study, we functionally characterized 25 variants identified during diagnostic testing or in exome sequencing studies. We assessed their kinetic characteristics, stability, and interaction with pharmacological and physiological regulators. We integrated our functional data with existing data from the ClinGen Monogenic Diabetes Variant Curation Expert Review panel using a gene-specific framework to assist variant classification. We show how functional evidence can aid variant classification, thus enabling diagnostic certainty. Full article

Background/Objectives: Artificial intelligence is revolutionizing healthcare across multiple areas, and periodontology is no exception to this emerging trend. This systematic study sought to rigorously assess the applicability and efficacy of artificial intelligence (AI) models in the diagnosis, classification, and prediction of periodontal diseases. Methods: A web-based search was performed across many reputable databases, including PubMed, Scopus, Embase, Cochrane, Web of Science, Google Scholar, and the Saudi Digital Library. Articles published between January 2000 and January 2025 were included in the search. Following the application of the inclusion criteria, 33 publications were selected for critical analysis utilizing QUADAS-2, and their certainty of evidence was evaluated using the GRADE technique. Results: The primary applications of AI technology include the diagnosis, classification, and grading of periodontal diseases; diagnosis of gingivitis; evaluation of the radiographic alveolar bone level and degree of alveolar bone loss; and prediction of periodontal disease risk. The AI models utilized in these studies outperformed current clinical methods in diagnosing, classifying, and predicting periodontal diseases, demonstrating a superior level of precision and accuracy. Their accuracies ranged from 73% to 99.4%, their sensitivities from 75% to 100%, and their precisions from 56% to 99.5%. Conclusions: AI has a lot of potential to help with periodontal diagnosis and risk assessment. Its performance is often similar to or better than that of traditional clinical approaches. But before it can be used widely in clinical settings, problems with the quality of the dataset, its generalizability, its interpretability, and its acceptance by regulators must be solved. AI should be seen as a tool that helps doctors make better decisions and not as a way to replace their knowledge and skills. Full article

Background: Microbiological culture and histology are gold standards for diagnosing chronic osteomyelitis (cOM) and fracture-related infection (FRI). This study investigated whether combining these modalities within a single tissue sample provides additional insight into disease severity. We hypothesized that high neutrophil and osteoclast numbers correlate with culture-positive microbiology and that double-positive samples may indicate more severe disease. Methods: In this prospective single-centre study, adults undergoing surgery for confirmed FRI or cOM were included. Clinical and disease classification data (FRI and BACH) were recorded. Five deep-tissue samples were collected intraoperatively and divided for paired microbiological culture and histological assessment of neutrophil infiltration, according to international diagnostic guidelines. Results: Forty-one patients were included (11 cOM, 30 FRI) of whom 68% received preoperative antibiotics. Nineteen patients (46%) were identified as culture-positive, while 32 patients (78%) were histologically positive according to international diagnostic guidelines, respectively. Among the 205 samples, 31% were culture-positive, 56% histology-positive, and 26% double-positive. Histological scores were significantly higher in culture-positive samples (p < 0.001). Treatment failure occurred in seven patients (18%), all with FRI. Paired positive samples were associated with increased odds of clinical failure and earlier revision, with odds increasing 1.68-fold for each additional paired positive sample (95% CI, 1.10–2.77). Conclusions: The paired analysis demonstrated a strong concordance between culture-positivity and suppurative inflammation within the same sample. Combining microbiology and histology may help identify patients at increased risk of revision and enhance diagnostic certainty, particularly in patients identified as culture-negative. Full article

Background: Vision-language models show promise in medical image interpretation, but their performance in musculoskeletal tumor diagnostics remains underexplored. Objective: To evaluate the diagnostic accuracy of six large language models on orthopedic radiographs for tumor detection, classification, anatomical localization, and X-ray view interpretation, and to assess the impact of demographic context and self-reported certainty. Methods: We retrospectively evaluated six VLMs on 3746 expert-annotated orthopedic radiographs from the Bone Tumor X-ray Radiograph dataset. Each image was analyzed by all models with and without patient age and sex using a standardized prompting scheme across four predefined tasks. Results: Over 48,000 predictions were analyzed. Tumor detection accuracy ranged from 59.9–73.5%, with the Gemini Ensemble achieving the highest F1 score (0.723) and recall (0.822). Benign/malignant classification reached up to 85.2% accuracy; tumor type identification 24.6–55.7%; body region identification 97.4%; and view classification 82.8%. Demographic data improved tumor detection accuracy (+1.8%, p < 0.001) but had no significant effect on other tasks. Certainty scores were weakly correlated with correctness, with Gemini Pro highest (r = 0.089). Conclusion: VLMs show strong potential for basic musculoskeletal radiograph interpretation without task-specific training but remain less accurate than specialized deep learning models for complex classification. Limited calibration, interpretability, and contextual reasoning must be addressed before clinical use. This is the first systematic assessment of image-based diagnosis and self-assessment in LLMs using a real-world radiology dataset. Full article

Background and Objectives: Artificial intelligence (AI) has shown promising performance in skin-lesion classification; however, its fairness, external validity, and real-world reliability remain uncertain. This systematic review and meta-analysis evaluated the diagnostic accuracy, equity, and generalizability of AI-based dermatology systems across diverse imaging modalities and clinical settings. Materials and Methods: A comprehensive search of PubMed, Embase, Web of Science, and ClinicalTrials.gov (inception–31 October 2025) identified diagnostic accuracy studies using clinical, dermoscopic, or smartphone images. Eighteen studies (11 melanoma-focused; 7 mixed benign–malignant) met inclusion criteria. Six studies provided complete 2 × 2 contingency data for bivariate Reitsma HSROC modeling, while seven reported AUROC values with extractable variance. Risk of bias was assessed using QUADAS-2, and evidence certainty was graded using GRADE. Results: Across more than 70,000 test images, pooled sensitivity and specificity were 0.91 (95% CI 0.74–0.97) and 0.64 (95% CI 0.47–0.78), respectively, corresponding to an HSROC AUROC of 0.88 (95% CI 0.84–0.92). The AUROC-only meta-analysis yielded a similar pooled AUROC of 0.88 (95% CI 0.87–0.90). Diagnostic performance was highest in specialist settings (AUROC 0.90), followed by community care (0.85) and smartphone environments (0.81). Notably, performance was lower in darker skin tones (Fitzpatrick IV–VI: AUROC 0.82) compared with lighter skin tones (I–III: 0.89), indicating persistent fairness gaps. Conclusions: AI-based dermatology systems achieve high diagnostic accuracy but demonstrate reduced performance in darker skin tones and non-specialist environments. These findings emphasize the need for diverse training datasets, skin-tone–stratified reporting, and rigorous external validation before broad clinical deployment. Full article

Over the past decade, recurring influxes of pelagic Sargassum have posed significant environmental and economic challenges in the Caribbean Sea. Effective monitoring is crucial for understanding bloom dynamics and mitigating their impacts. This study presents a comprehensive machine learning (ML) and deep learning (DL) framework for detecting Sargassum and estimating its fractional cover using imagery from key satellite sensors: the Operational Land Imager (OLI) on Landsat-8 and the Multispectral Instrument (MSI) on Sentinel-2. A spectral library was constructed from five core spectral bands (Blue, Green, Red, Near-Infrared, and Short-Wave Infrared). It was used to train an ensemble of five diverse classifiers: Random Forest (RF), K-Nearest Neighbors (KNN), XGBoost (XGB), a Multi-Layer Perceptron (MLP), and a 1D Convolutional Neural Network (1D-CNN). All models achieved high classification performance on a held-out test set, with weighted F1-scores exceeding 0.976. The probabilistic outputs from these classifiers were then leveraged as a direct proxy for the sub-pixel fractional cover of Sargassum. Critically, an inter-algorithm agreement analysis revealed that detections on real-world imagery are typically either of very high (unanimous) or very low (contentious) confidence, highlighting the diagnostic power of the ensemble approach. The resulting framework provides a robust and quantitative pathway for generating confidence-aware estimates of Sargassum distribution. This work supports efforts to manage these harmful algal blooms by providing vital information on detection certainty, while underscoring the critical need to empirically validate fractional cover proxies against in situ or UAV measurements. Full article

Earthquakes are sudden and destructive natural events caused by tectonic movements in the Earth’s crust. Although they cannot be predicted with certainty, rapid and reliable detection is essential to reduce loss of life and property. This study aims to automatically distinguish earthquake and noise signals from real seismic data by analyzing time-frequency features. Signals were scaled using z-score normalization, and extracted with Empirical Mode Decomposition (EMD), Discrete Wavelet Transform (DWT), and combined EMD+DWT methods. Feature selection methods such as Lasso, ReliefF, and Student’s t-test were applied to identify the most discriminative features. Classification was performed with Ensemble Bagged Trees, Decision Trees, Random Forest, k-Nearest Neighbors (k-NN), and Support Vector Machines (SVM). The highest performance was achieved using the RF classifier with the Lasso-based EMD+DWT feature set, reaching 100% accuracy, specificity, and sensitivity. Overall, DWT and EMD+DWT features yielded higher performance than EMD alone. While k-NN and SVM were less effective, tree-based methods achieved superior results. Moreover, Lasso and ReliefF outperformed Student’s t-test. These findings show that time-frequency-based features are crucial for separating earthquake signals from noise and provide a basis for improving real-time detection. The study contributes to the academic literature and holds significant potential for integration into early warning and earthquake monitoring systems. Full article

The southwestern region of China is highly susceptible to landslides due to steep terrain, fractured geology, and intense rainfall. This study focuses on the Xiluodu Reservoir area in Yunnan Province and applies Geographic Information System (GIS) techniques together with ten key spatial factors—such as slope, lithology, elevation, and distance to rivers—to perform a quantitative landslide risk assessment. In addition to the individual Certainty Factor (CF) and Logistic Regression (LR) models, we developed an integrated CF–LR coupled model to overcome their respective limitations: the CF model’s sensitivity to specific factor attributes but neglect of factor interactions, and the LR model’s robust weight estimation but weak representation of attribute heterogeneity. By combining these strengths, the CF–LR model achieved superior predictive performance (AUC = 0.804), successfully capturing 92.5% of historical landslide events within moderate-to-high risk zones. The results show that lithology, slope angle, and proximity to rivers and roads are dominant controls on susceptibility, with landslides concentrated on soft rock slopes of 30–40° and within 600–900 m of rivers. Compared with previous coupled approaches in similar mountainous reservoir settings, our CF–LR model provides a more balanced and interpretable framework, enhancing both classification accuracy and practical applicability. These findings demonstrate that GIS-based CF–LR integration is a novel and reliable tool for landslide susceptibility mapping, offering important technical support for disaster prevention and risk management in large reservoir regions. Full article

Objectives: Pulmonary embolism (PE) is an uncommon but potentially fatal complication of acute ischaemic stroke (AIS). Its global burden and prevention remain incompletely defined. We performed a systematic review and meta-analysis (PEARL-AIS) to estimate prevalence, risk factors, outcomes, and prophylactic efficacy, with GRADE evidence appraisal. Methods: Following PRISMA 2020 and MOOSE guidelines, five databases (PubMed, Embase, Cochrane, Scopus, Web of Science) were searched (1995–2024). The protocol was prospectively registered (OSF s25ny). Random-effects models (DerSimonian–Laird; REML sensitivity) were used to pool prevalence and odds ratios; heterogeneity was evaluated with I², Cochran’s Q, and τ². Influence (leave-one-out) and subgroup analyses for prevalence and mortality of PE in AIS were explored. Bias was assessed using the Modified Jadad Scale; overall certainty was graded with the GRADE framework. Results: Twenty-four studies met the inclusion criteria (n = 25,666,067), of which seventeen studies (n = 23,637,708) contributed to pooled prevalence analyses. The pooled prevalence of PE after AIS was 0.40% (95% CI 0.33–0.49), approximately six-fold higher than in the general population, with considerable heterogeneity (I² > 90%, Cochrane classification). The pooled mortality among AIS patients with PE was 12.9% (95% CI 1.6–31.7). Mortality risk was significantly higher in AIS patients with PE (OR 4.96, 95% CI 2.98–8.24). Atrial fibrillation (29%), cancer (19%), and smoking (23%) were common; hypertension (54%) and diabetes (23%) were prevalent but not predictive, with diabetes showing a paradoxical protective association (OR 0.88, 95% CI 0.84–0.92). Pharmacological prophylaxis was associated with a reduced risk of PE (OR 0.64, 95% CI 0.46–0.90; I² = 0%), supported by moderate-certainty evidence. Conclusions: PE is an uncommon but often fatal complication of AIS. Traditional venous thromboembolism predictors underperform in this context, suggesting a stroke-specific thromboinflammatory mechanism linking the brain and lung axis. Despite considerable heterogeneity and low-to-moderate certainty of evidence, pharmacological prophylaxis demonstrates a consistent protective effect. Systematic PE surveillance and tailored prophylactic strategies should be integral to contemporary stroke care, while future studies should refine risk stratification and elucidate the mechanistic underpinnings of this brain–lung thromboinflammatory continuum. Full article