Search Results (3,365)

This study evaluates the effects of a 30-week mindfulness-enhanced resistance training (MRT) program on the physical and mental health of female college students and explores whether changes in self-esteem or mindfulness mediate the relationship between MRT and subjective well-being. Sixty-four healthy female college students were randomly assigned to either the MRT or resistance training (RT) group. Both groups participated in 90 min weekly sessions for 30 weeks. A 2 × 2 mixed-design ANOVA analyzed the intervention’s effects on physical health, mindfulness, self-esteem, and subjective well-being. PROCESS macro (Model 4) tested mediation effects. MRT and RT significantly improved physical health, with MRT showing superior improvements in waist-to-hip ratio, flexibility, and vital capacity. Only MRT improved mindfulness, self-esteem, and subjective well-being. Self-esteem changes fully mediated the relationship between MRT and subjective well-being. MRT as a comprehensive mind–body intervention significantly enhanced the physical health and subjective well-being of female college students, outperforming resistance training. Improvements in self-esteem mediated the relationship between MRT and increased subjective well-being. MRT can serve as an effective approach to promote the physical and mental health of female college students. Full article

Cross-topic stance detection degrades when encoders entangle stance signals with topic-specific vocabulary, causing representations that fail to transfer to unseen targets. Existing methods commit to either topic-invariant or topic-aware representations and apply the same strategy uniformly to every input, sacrificing complementary information. We propose ReDyGait, a three-stage framework that disentangles these two types of signals through dedicated contrastive pre-training and recombines them adaptively at inference time. Stage 1 trains a topic-invariant encoder with supervised contrastive loss over cross-topic positives. Stage 2 trains a topic-contextual encoder with bidirectional pair contrastive loss over within-topic positives; both stages employ topic-aware hard negative mining to prevent shortcut learning. Stage 3 freezes the two contrastive encoders and learns a gating network that produces per-instance weights over invariant, contextual, and base-encoder pathways. On VAST, ReDyGait achieves a macro-averaged F1 of 0.782 in the zero-shot setting and 0.752 in the few-shot setting, improving over the strongest baseline by 1.1 points in both; on SEM16t6 in a leave-one-target-out setup, ReDyGait reaches an average F1 of 0.612. Analysis of the learned gate weights shows that the model shifts toward the invariant pathway for unfamiliar topics and toward the contextual pathway when topic-specific patterns are available, confirming that the disentanglement operates as intended. Full article

Within a waste biorefinery framework, integrating agro-industrial by-products into the circular economy requires a detailed understanding of the thermochemical conversion behaviour of low-grade carbonaceous materials. This study evaluates the co-pyrolysis characteristics of Soma lignite (SL) and pectin-rich sugar beet pulp (SBP) as a sustainable route for upgrading these resources into clean energy carriers. Interactions between the two feedstocks were analysed by thermogravimetric measurements, triple-region kinetic modelling, and quantitative synergy indices at six mixing ratios, including the pure samples (100:0, 80:20, 60:40, 40:60, 20:80, and 0:100 wt% SL:SBP). The Reactivity Index ($R_{\mathrm{m}}$) increased from $0.97\times {10}^{-4}$ s${}^{-1}$K${}^{-1}$ for pure SL to $8.65\times {10}^{-4}$ s${}^{-1}$K${}^{-1}$ for the 20:80 blend, showing that SBP acts as a highly reactive biomass component that accelerates devolatilisation in the main pyrolysis region. Synergy analysis indicated a shift from inhibitory behaviour in coal-rich blends to slightly positive synergy in SBP-rich mixtures, with the onset of positive $\Delta TC$ around 60 wt% SBP under the present single-heating-rate, non-replicated TGA conditions. This tentative threshold-like behaviour suggests that a critical level of literature-supported, hypothesised hydrogen-donating biomass radicals may be required to overcome the structural resistance of the coal matrix. Within these experimental limitations, the apparent macro-kinetic deviations and first-order Arrhenius parameters suggest that SL/SBP co-pyrolysis follows a complex, non-additive pathway that should be further validated by multi-heating-rate and product characterisation studies in future work. The primary contribution of this work lies in proposing this distinct threshold-like biomass fraction at the macro-kinetic level that governs the transition from heat-transfer-limited antagonism to radical-influenced synergy in low-rank coal and pectin-rich biomass blends. Overall, the combined $\Delta TC$, $\Delta E$ and $R_{\mathrm{m}}$ descriptors provide useful macro-kinetic benchmarks for guiding the optimisation of thermochemical processes for low-grade carbonaceous resources. Full article

Predictive maintenance plays a key role in digitalization initiatives; however, in real settings, issues related to failure prediction occur when failure instances are rare compared to normal instances, leading to class imbalance. In this study, we systematically compare five machine learning (ML) models—random forest, XGBoost, support vector machine, k-nearest neighbors, and multinomial logistic regression (MLR)—to detect multiclass rare failures using four imbalance-handling approaches (i.e., no handling, manual oversampling, selective manual oversampling, and class weighting), forming 20 configurations. Using the AI4I 2020 predictive maintenance dataset, which contains five failure types, we determined that XGBoost with no handling achieved the highest macro-averaged F1 (macro-F1) score (0.842) but obtained 0% recall for tool wear failure (TWF). MLR with selective manual oversampling achieved approximately 50% TWF recall with lower overall performance (0.636 macro-F1) than top-performing models such as XGBoost. We also found that very rare classes remain difficult to detect. Even high-performing models fail to consistently detect all five failure types. Overall, no single strategy can achieve a high detection rate across all performance measures. Full article

Analyzing large-scale student feedback is critical for higher education quality assurance, yet manual analysis is inefficient and subjective. This paper proposes an integrated framework that unifies supervised classification, unsupervised clustering, and retrieval-augmented generation (RAG) to produce evidence-grounded and actionable insights. Ensemble-based supervised models perform thematic classification, while multi-encoder embedding fusion enables unsupervised discovery of coherent feedback clusters. A multi-stage RAG module integrates category predictions and cluster structure to retrieve representative evidence and generate transparent summaries with citation traceability. The framework is evaluated on student feedback collected from a Central Asian university and two public benchmarks, EduRABSA and Coursera course reviews, covering seven thematic categories. The supervised ensemble achieves 83.0% accuracy and 0.829 Macro-F1 on the primary dataset, while unsupervised clustering attains a silhouette score of 0.271 under the best fusion strategy. Independent evaluation on external benchmarks yields ensemble accuracy of 81.1% on EduRABSA and 49.8% on Coursera, confirming the framework’s adaptability across diverse educational contexts. By leveraging supervised labels and unsupervised structure, the proposed framework enables evidence-grounded, category-aware LLM-based summaries that faithfully reflect the diversity and distribution of student feedback and support actionable educational decision-making. Full article

Against the backdrop of a multi-tiered governance system and increasingly institutionalized norms, China’s historical and cultural preservation policies have long emphasized institutional standardization and hierarchical uniformity. Local policy texts are typically viewed as localized replicas of central institutional logic, overlooking internal variations and differences in information structure. Accordingly, this study examines the Regulations on the Protection of Historical and Cultural Cities, Towns, and Villages issued by 13 provincial-level administrative regions in China. It conceptualizes provincial regulatory texts as institutionalized policy information systems, constructs a cross-regional corpus, and develops a comparative information structure analytical framework based on the Latent Dirichlet Allocation (LDA) topic model. This study operationalizes LDA-derived topic-weight distributions into a comparative analytical framework that captures structural prominence, dispersion, concentration, and priority hierarchy in provincial policy texts. The findings reveal that provincial-level historical and cultural preservation regulations in China exhibit a highly institutionalized information backbone, centered on administrative procedures, legal norms, and macro-level planning controls, and demonstrate significant institutional similarity across provinces. However, within this unified institutional framework, provinces exhibit structural differences in the distribution of thematic weights, information prioritization, and internal textual sequencing, resulting in multiple distinguishable information organization patterns. Consequently, this study highlights the coexistence of formal institutional uniformity and structural differentiation in provincial regulatory texts, providing a more precise basis for understanding variation in local policy expression within China’s historical and cultural governance field. Full article

The uneven spatial distribution of data elements poses challenges to regional equity and sustainable development. To unmask spatial dynamics obscured by traditional macro-divisions, this study evaluates data element development across China’s Eight Comprehensive Economic Regions from 2013 to 2022. Using the entropy weight method, Dagum Gini coefficient, Kernel Density Estimation, and spatial autocorrelation models, the results indicate that while the overall development index exhibits a sustained upward trend, inter-regional differences remain the dominant source of spatial inequality. This disparity is primarily driven by the persistent gap between advanced coastal and lagging inland regions. Notably, spatial trajectories diverge significantly: the Eastern Coastal region exhibits coordinated integration, whereas severe internal polarization appears in the Middle Reaches of the Yellow River and the Southwest. Furthermore, the spatial spillover of data elements remains bounded by physical geography. By highlighting these meso-level structural fault lines, this study provides precise empirical evidence for formulating targeted, basin-specific interventions to bridge the digital divide. Full article

Real-time facial emotion recognition remains challenging due to the high dimensionality and computational cost of dense facial representations, which limit their applicability in resource-constrained and real-time scenarios. This study proposes a compact, anatomically informed angular facial representation for efficient, interpretable emotion recognition under real-time constraints. Facial landmarks are first extracted using a standard landmark detection framework, from which a reduced facial mesh of 27 anatomically selected points is defined. Internal geometric angles computed from this mesh are analyzed using temporal variability and redundancy criteria, resulting in a minimal set of eight angular descriptors that capture the most expressive facial dynamics while preserving geometric invariance and computational efficiency. The proposed representation is evaluated using multiple supervised machine learning classifiers under two complementary validation strategies: stratified frame-level cross-validation and strict Leave-One-Subject-Out evaluation. Under mixed-subject stratified validation, the best-performing model (MLP) achieved macro-averaged F1-scores exceeding 0.95 and near-unity ROC–AUC values. However, subject-independent evaluation revealed reduced generalization performance (average accuracy ≈55%), highlighting the influence of inter-subject morphological variability embedded in absolute angular descriptors. These findings indicate that a minimal angular geometric encoding provides strong intra-subject discriminative capability while transparently characterizing its cross-subject generalization limits, offering a practical and interpretable alternative for data- and resource-constrained real-time scenarios. Full article

Introduction: Circadian misalignment has been proposed as a potential determinant of cardiometabolic risk. Chronotype, as an expression of individual circadian organization, has been associated with unfavorable metabolic profiles; however, the role of total and central adiposity as potential mediating mechanisms in this relationship remains incompletely understood. Objective: This study aimed to analyze the association between chronotype and cardiometabolic risk in adults and to evaluate the potential mediating role of body fat percentage (BF%) and waist-to-height ratio (WHtR). Methods: An observational study was conducted in 1462 adults from the general population. Chronotype was assessed using the Morningness–Eveningness Questionnaire (MEQ), and cardiometabolic risk was evaluated using a continuous cardiometabolic risk score (CMRS) derived from waist circumference (WC), systolic blood pressure (SBP), triglycerides (TG), fasting blood glucose (FBG), and total cholesterol (TC). Multiple linear regression models adjusted for covariates were used to examine the association between chronotype and CMRS, and hierarchical regression was performed to estimate the incremental contribution of adiposity indicators. Mediation analysis was conducted using the PROCESS macro (Model 4) with 95% bootstrap confidence intervals. Results: Chronotype was independently associated with CMRS after adjustment for covariates (β = 0.055; p = 0.030), although the effect size and explained variance were small. In hierarchical regression analysis, the inclusion of chronotype explained a small but significant increase in CMRS variance (ΔR² = 0.003; p = 0.030). The addition of adiposity indicators significantly increased the explained variance (ΔR² = 0.014; p < 0.001), with WHtR emerging as the most relevant predictor in the final model. Bootstrap mediation analysis did not reveal significant indirect effects of BF% or WHtR on the relationship between chronotype and CMRS. In sensitivity analyses excluding waist circumference from the CMRS, the association between chronotype and cardiometabolic risk was no longer significant (β = −0.001; p = 0.974). Conclusions: Chronotype showed a modest association with cardiometabolic risk in the primary analysis. However, sensitivity analyses indicated that this association may partly depend on the inclusion of waist circumference within the composite cardiometabolic risk score. These findings highlight the central role of abdominal adiposity in cardiometabolic health and suggest that the relationship between chronotype and cardiometabolic risk should be interpreted with caution. Full article

This paper investigates the quality of mobile network coverage along the Riga–Tukums railway corridor with a focus on the performance of 4G and 5G technologies. Ensuring reliable mobile connectivity along suburban railway corridors remains a significant technical challenge due to mixed forest–urban propagation conditions, macro-cell-dominated LTE infrastructure, mobility-induced channel variability, and fluctuating passenger density. Unlike high-speed railway environments that are extensively studied in dedicated 5G-R scenarios, suburban railway systems often rely on existing macro-cell deployments, where coverage continuity, signal quality stability, and capacity constraints must be addressed simultaneously. This study presents a measurement-based evaluation of 4G and 5G radio performance along the Riga–Tukums railway corridor under real operational conditions (50–90 km/h). Classical propagation models (Okumura–Hata and COST231-Hata) are quantitatively validated using MAE and RMSE metrics, followed by correlation analysis between RSSNR and QoS indicators. A theoretical Doppler sensitivity assessment (80–200 km/h) is conducted to evaluate mobility robustness across LTE and 5G frequency bands. Mobility transition regions and handover-related time windows are geometrically estimated, and passenger density-based capacity modeling is applied to assess throughput degradation under peak occupancy scenarios. Based on these results, a multi-layer network planning strategy integrating 700 MHz macro coverage, 1700 MHz capacity enhancement, and 3500 MHz 5G NR deployment is proposed. The optimization strategy resulted in an estimated 22–28% increase in stable service coverage in previously weak-signal zones and demonstrated that propagation model deviations remain within ranges comparable to recent railway studies (≈15–25 dB RMSE). These findings provide a structured framework for suburban railway communication optimization and support the gradual modernization of railway infrastructure toward FRMCS-ready architectures. The study illustrates the applicability of modern modelling tools for assessing and improving mobile communication systems and contributes to the broader development of digital infrastructure within Latvia’s transport sector. Full article

Understanding the mechanical behavior of lunar regolith is critical for the success of future lunar excavation and construction missions. Irregular particle morphology and low geostatic stress are recognized as key factors contributing to the high internal friction angle of this unique extraterrestrial geomaterial. However, the underlying mechanisms by which low geostatic stress enhances shear strength remain unclear, and the multiscale effects of particle morphology on shear strength evolution are not yet fully elucidated. In this study, consolidated drained triaxial compression tests were performed on CUMT-1 lunar regolith simulant and Fujian standard sand to investigate their macroscopic mechanical behavior. Complementary discrete element simulations of biaxial compression were conducted to analyze mesoscopic mechanical responses of granular materials under the influence of multiscale particle morphology and confining stress. A robust macroscopic–mesoscopic strength correlation model was established, incorporating normalized mean interparticle contact force and mean coordination number to predict the normalized deviatoric stress of granular assemblies. Based on this model, the mesoscopic mechanisms through which irregular particle morphology and low geostatic stress enhance the internal friction angle were quantitatively investigated. The findings offer new insights into the shear strength characteristics of in situ lunar regolith and provide theoretical support for lunar surface construction and excavation operations. Full article

In UAV (unmanned aerial vehicle)-enabled precision agriculture, achieving high-accuracy disease diagnosis and yield estimation simultaneously on resource-constrained edge devices remains a significant challenge. Existing solutions are commonly hindered by conflicts in visual feature scales, the absence of explicit agronomic causal logic, and the trade-off between lightweight design and global modeling capability. To address these challenges, a heterogeneous dual-stream state-space framework termed SymbioMamba is proposed. The proposed framework incorporates three key innovations: first, a heterogeneous dual-stream encoder is constructed, in which a micro-texture stream captures high-frequency disease details while a macro-context-scan stream models field-scale biomass continuity; second, a pathology–biomass collaborative interaction (PBCI) module is designed to explicitly inject the biological prior—disease stress leading to yield reduction—into the feature space. Third, a topology-aligning cross-architecture distillation (TACAD) paradigm is introduced to transfer global knowledge from a heavyweight teacher to a lightweight student. Experimental results from a maize UAV dataset comprising 12,074 annotated image patches demonstrate that SymbioMamba achieves 89.4% mAP@0.5 and an $R^2$ of 0.915. Compared to the industry-standard YOLOv11, the framework improves mAP@0.5:0.95 by 2.4% while reducing the parameter count to 6.2 M—a 50% decrease relative to monolithic state-space baselines. Furthermore, yield prediction error is significantly reduced to an RMSE of 485.6 kg/ha. With a compact model size of 6.2 M parameters and 2.4 G FLOPs, SymbioMamba attains an inference speed of 38.2 FPS on the NVIDIA Jetson AGX Orin platform, providing a high-performance, real-time solution for intelligent agricultural phenotypic analysis. Full article

Green innovation is central to industrial ecological transition, yet heavily polluting firms often exhibit low-quality green innovation investment. Grounded in the government–enterprise–research–intermediary Quadruple Helix innovation ecosystem framework, this study integrates Necessary Condition Analysis (NCA) and fuzzy set qualitative comparative analysis (fsQCA) to examine 66 publicly listed heavily polluting manufacturing firms in China. The results show that fiscal subsidies and environmental taxes are necessary but not sufficient conditions for achieving high-quality green innovation investment. Moreover, high-quality outcomes arise through three equifinal pathways: the Government–Intermediary Dual-Drive Model, the Government–Enterprise–Intermediary Co-Directional Model, and the Government–Enterprise Symbiotic Model. Six configurations lead to non-high-quality green innovation investment, which cluster into Resource-Scarcity and Regulatory-Constrained models. A favorable macro environment further strengthens high-quality outcomes. These findings clarify how policy instruments and multi-actor collaboration jointly shape green innovation investment quality and provide actionable implications for heavily polluting firms and policymakers seeking sustainable development. Full article

The increasing heterogeneity of cloud, enterprise, and Internet of Things (IoT) environments raises concerns about the long-term reliability of machine-learning-based intrusion detection systems (IDSs). This study evaluates temporal robustness and cross-domain generalization using four publicly available datasets collected between 2017 and 2024. Five representative models—Random Forest, Gradient Boosting, Multi-Layer Perceptron, Autoencoder, and a lightweight 1D-CNN—are assessed under in-dataset, forward temporal, enterprise-to-IoT transfer, and dataset-agnostic evaluation protocols without retraining. In the dataset evaluation, models achieve Macro-F1 scores between 0.84 and 0.96. However, forward temporal testing reveals consistent degradation, with performance reductions reaching ΔF1 ≈ 0.20–0.27 when models trained on 2017 enterprise traffic are applied to IoT datasets from 2023 to 2024. Under cross-domain transfer, Macro-F1 decreases to 0.69–0.78, and benign false-positive rates increase up to 0.30, indicating substantial sensitivity to traffic distribution shifts. Tree-based ensemble models show comparatively lower degradation (≈6–23%) and reduced performance variance across datasets. Semantic feature analysis further indicates that flow intensity and temporal activity features exhibit higher cross-dataset stability than protocol-dependent indicators. These findings demonstrate that IDS robustness in evolving Internet environments depends strongly on evaluation methodology and feature stability, highlighting the need for generalization-oriented assessment strategies. Full article

Image classification in low-data regimes remains a challenging problem, particularly in stylized visual domains where intra-class similarity and inter-class feature overlap limit discriminative capacity. This study presents a systematic evaluation of regularization and transfer learning strategies for multi-class comic character recognition under constrained data conditions. Four convolutional architectures are compared: (i) a baseline CNN trained from scratch, (ii) a regularized CNN incorporating data augmentation, dropout, and early stopping, (iii) a pretrained ResNet-50 used as a fixed feature extractor, and (iv) a partially fine-tuned ResNet-50 with selective layer unfreezing. Experiments are conducted on a custom four-class dataset exhibiting moderate class imbalance, evaluated using both a fixed 70/20/10 split and 5-fold cross-validation to assess generalization stability. Results indicate that shallow CNN architectures suffer from substantial overfitting, even when regularization is applied, whereas transfer learning significantly improves macro-averaged F1-score and out-of-distribution detection performance. Cross-validated results, the primary basis for inference given the dataset scale, show that both ResNet-50 strategies achieve equivalent mean accuracy of 95.0% (SD: ±0.4% for feature extraction, ±0.8% for fine-tuning; paired t = 0.00, p = 1.000), while shallow CNN architectures reach only 81–87%. Under a single fixed 70/20/10 partition (n = 69 test samples, 95% CI: ±9–12%), fine-tuning nominally reaches 98.5%; crucially, cross-validation deflates this figure to parity with feature extraction, confirming it reflects favorable partitioning rather than genuine architectural superiority. The primary finding is therefore that frozen ResNet-50 feature extraction is the recommended strategy: it matches fine-tuning in cross-validated generalization while requiring 15× fewer trainable parameters and exhibiting lower fold-to-fold variance. The findings demonstrate that pretrained deep residual representations transfer effectively to stylized comic imagery and that evaluation protocol selection critically impacts perceived performance in small datasets. These results provide practical guidelines for robust model selection in domain-specific, limited-data image classification tasks. Full article