Search Results (192)

Cattle behavior constitutes important phenotypic information reflecting animals’ health status, activity level, and welfare condition, and is therefore of considerable significance for automated monitoring and precision management in smart livestock farming. However, under complex barn conditions, cattle behavior recognition is easily affected by factors such as illumination variation, partial occlusion, background interference, and individual differences, thereby reducing recognition stability and generalization capability. To address these challenges, this study proposes a pose-driven method for cattle behavior recognition in complex barn environments. First, a 16-keypoint annotation scheme suitable for describing bovine posture, termed cow16, was constructed. Based on this scheme, OpenPose was employed to extract heatmaps (HMs) and part affinity fields (PAFs), which were then used to build an intermediate HM/PAF posture representation. Subsequently, this representation was taken as the input to a lightweight convolutional neural network for classifying three behavioral categories: stand, walk, and lying. On this basis, class-imbalance correction during training and a multi-random-seed logits ensemble strategy during inference were further introduced. In addition, knowledge distillation was adopted to transfer knowledge from a high-performance teacher model to a lightweight student model. Experimental results demonstrate that training-stage class-imbalance correction and inference-stage multi-random-seed logits ensembling exhibit strong complementarity; when combined, the AB configuration improves the test-set Macro-F1 by 3.83 percentage points. Moreover, the distilled student model still achieves competitive recognition performance while maintaining 1× inference cost, indicating a favorable trade-off between accuracy and efficiency. This study provides a useful reference for deployment-oriented cattle behavior recognition in smart farming scenarios and offers a lightweight technical basis for subsequent practical applications. Full article

Financial time series forecasting remains a persistent challenge due to the non-stationary nature, inherent noise, and multi-scale temporal dependencies present in market data. This paper presents MSTFNet, a multi-scale temporal fusion network that combines dilated causal convolutions with a frequency-enhanced sparse attention mechanism for improved financial prediction. The proposed architecture consists of three core components: a multi-scale dilated causal convolution module that extracts temporal patterns across different time horizons through parallel convolutional branches with varying dilation rates, a frequency-enhanced sparse attention mechanism that leverages Fast Fourier Transform to identify dominant periodic components and modulate attention weights accordingly, and an adaptive scale fusion gate that learns to dynamically combine representations from multiple temporal scales. Extensive experiments conducted on three public financial datasets (S&P 500, CSI 300, and NASDAQ Composite) spanning the period from January 2015 to December 2024 show two key results. First, consistent with near-efficient markets, the random-walk benchmark (${\widehat{y}}_{t+1}=y_t$) outperforms all the data-driven models on level-error metrics (MAE, RMSE, MAPE, and $R^2$), establishing the martingale as the binding lower bound on point-prediction error. Second, MSTFNet achieves the highest directional accuracy (DA) across all three indices—56.3% on the S&P 500 versus 50.0% for the martingale—representing a 6.3 percentage-point improvement that generates positive pre-cost returns in a trading strategy backtest. Among the eight data-driven baselines (LSTM, GRU, TCN, Transformer, Autoformer, FEDformer, PatchTST, and iTransformer), MSTFNet also achieves the lowest MAE, reducing it by 13.6% relative to the strongest data-driven baseline (iTransformer) on the S&P 500. These results confirm that integrating multi-scale temporal modeling with frequency-domain guidance extracts a real, if modest, directional signal from financial time series. Full article

Background/Objectives: Emerging evidence shows that dual tasking as well as the restriction of arm movements independently lead to detrimental effects on walking performance. However, it is unclear whether the deteriorations are more pronounced when applied together and if children (i.e., due to ongoing maturation processes) perform differently compared to young adults. This study investigated the influence of different arm movement strategies on subjective and objective markers related to beam walking under single-task (ST) and dual-task (DT) conditions in children and young adults. Methods: Twenty-six children (age: 11.3 ± 0.6 years) and 30 young adults (age: 23.2 ± 2.8 years) walked three meters on a balance beam with free and restricted (i.e., arms crossed over the chest) arm movements in a random order while concurrently performing a cognitive task (i.e., serial subtractions) or not. Walking outcomes (i.e., gait speed, cadence) were measured and used as objective markers. Self-reported task-related perceptions (i.e., balance confidence, fear of falling, perceived instability, conscious balance processing) were assessed and used as subjective indicators. Results: Walking under DT conditions (i.e., main effects of task) detrimentally influenced subjective task-related perceptions and walking outcomes, but using free arm movements (i.e., task × arm interactions) mitigated these deteriorations. Further, children exhibited largely stable levels of conscious balance processing, whereas young adults demonstrated overall higher levels along with pronounced differences between ST and DT walking when arm movements were unrestricted (i.e., group × task × arm interaction). Conclusions: These findings indicate that free arm movements seem to constitute a simple yet effective complementary ‘upper-body strategy’ that enhances postural control during a cognitively demanding walking task. Further, age differences imply that young adults compensate demanding walking conditions (i.e., DT walking with restricted arms) by elevated conscious processing of balance (i.e., a shift from automated to more conscious attention towards postural control). Full article

Background: Pulmonary hypertension (PH) remains a progressive and life-threatening condition despite advances in targeted pharmacotherapy. Pulmonary artery denervation (PADN) has emerged as a novel interventional strategy aimed at modulating sympathetic overactivity and improving pulmonary vascular hemodynamics. Methods: A comprehensive search of PubMed, EMBASE, Scopus, Web of Science, and the Cochrane Library was conducted through December 2024. Randomized clinical trials and prospective observational studies assessing PADN in PH were included. Primary endpoints were changes in outcomes from six-minute walk distance (6MWD), mean pulmonary artery pressure (mPAP), pulmonary vascular resistance (PVR), cardiac output (CO), and right ventricular function parameters. Secondary outcomes included clinical worsening, rehospitalization, transplantation, and all-cause mortality. Random-effects models were used to calculate pooled mean differences (MDs) and odds ratios (ORs) with 95% confidence intervals (CIs). Subgroup analyses were performed according to pulmonary hypertension phenotype and study design, and sensitivity analyses were conducted to assess robustness of pooled estimates. Results: Nine studies involving 454 patients were included. PADN significantly improved functional capacity (6MWD: MD = 92.03 m; 95% CI 46.37–137.68; p < 0.001) and reduced mPAP (MD = −11.84 mmHg; p < 0.001) and PVR (MD = −4.88; p < 0.001). Cardiac output increased significantly (MD = 0.55 L/min; p < 0.001), with improvements observed in right ventricular functional indices. PADN was associated with a lower risk of clinical worsening (OR = 0.30; p = 0.001) and rehospitalization (OR = 0.07; p < 0.001), whereas no significant difference was observed in all-cause mortality (OR = 0.53; p = 0.12). Considerable heterogeneity was observed across functional and hemodynamic outcomes, reflecting variability in study design, patient populations, and PADN techniques. Conclusions: PADN significantly improves exercise capacity and pulmonary hemodynamics in patients with PH, particularly in those with persistent symptoms despite medical therapy. Although PADN reduces clinical deterioration and rehospitalization, its impact on long-term survival remains uncertain. Further large-scale, multicenter randomized trials are needed to better define optimal patient selection and determine long-term clinical benefit. Full article

Soil nutrients and water are often distributed heterogeneously in space, yet how plant roots forage in response to such heterogeneity and how their strategies relate to functional traits remain poorly understood. Here, we conducted an indoor pot experiment manipulating water and nutrient supply in both homogeneous and heterogeneous patch patterns using seedlings of four tree species, focusing on root functional traits and foraging strategies. The results indicate that root foraging behavior exhibits both resource specificity and species specificity: roots tend to proliferate toward nutrient-rich and low-water patches as an adaptive strategy. Although no strict dichotomy was observed between high foraging scale (low precision) and low foraging scale (high precision) strategies under heterogeneous conditions, fine-rooted species (Acer truncatum and Koelreuteria paniculata) exhibited traits leaning toward “precise foraging”, whereas coarse-rooted species (Prunus davidiana and Quercus variabilis) tended toward a conservative “random walk” pattern, with no trade-off between root foraging scale and precision. Root morphological traits exerted significant nonlinear regulation on foraging scale: root biomass foraging scale (FS_RB) correlated positively with root diameter (RD) but negatively with specific root length (SRL) and specific root area (SRA); root length foraging scale (FS_RL) correlated positively with root length (RL), root tip number (RTN), SRL, and SRA. In contrast, root morphological traits could not explain the variation in foraging precision, suggesting that foraging precision constitutes another distinct dimension in root-trait space. In summary, this study provides key insights into the foraging strategies of plant roots in heterogeneous environments, expanding our understanding of the multidimensionality of root functional traits. Full article

The Green Oasis (GRO) Project is a targeted urban greening intervention designed to evaluate the environmental and health impacts of compact, high-density plantings in dense built environments. Initiated in downtown Louisville, the project transformed Founders Square, a 0.64-acre sparsely planted park, into a microforest (“Trager Microforest”), a multilayered planting of 119 trees and more than 200 shrubs. The impact of this intervention is being assessed through a randomized crossover study in which participants walk in the microforest and a nearby impervious parking lot. Physiological outcomes include heart rate, heart rate variability, arterial stiffness, and stress biomarkers measured in saliva, urine, and sweat. Environmental conditions are continuously monitored by fixed and mobile weather stations, air pollution sensors, and biodiversity surveys. Baseline assessments were conducted in 2023 and 2024, with post-planting evaluations now underway (2025–). Power calculations indicate adequate sensitivity (n ≈ 40–50) to detect changes in cardiovascular stress responses in participants. Complementary ecological measurements include soil microbiome composition, greenhouse gas fluxes, and avian diversity. This study addresses critical gaps in understanding how small-scale, high-density greening interventions affect cardiovascular resilience, stress physiology, and microclimatic regulation. By integrating environmental, biological, and human health data, GRO establishes a comprehensive framework for evaluating the efficacy of urban microforests as nature-based solutions. The results are expected to inform urban planning, public health strategies, and climate adaptation policies, demonstrating how compact greening interventions can simultaneously mitigate heat, reduce pollution, enhance biodiversity, and promote human wellbeing in dense urban cores. Full article

To enhance the accuracy of short-term photovoltaic (PV) power forecasting, this study proposes a novel hybrid model that integrates Random Forest (RF), Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN), Sample Entropy (SE), the Random Walk with Compulsory Evolution (RWCE) algorithm, and the Gated Recurrent Unit (GRU) network. Initially, RF is applied to select relevant meteorological features, minimizing redundancy and improving both training efficiency and predictive robustness under complex operating conditions. ICEEMDAN is then employed to decompose the PV power series into multiple quasi-stationary components, mitigating the adverse effects of non-stationarity on forecasting accuracy. Following this, SE is applied to quantify the complexity of each component and reconstruct the decomposed signals into high-, mid-, and low-frequency bands, simplifying the inputs to the forecasting model. To further improve performance, the RWCE algorithm optimizes GRU network hyperparameters through global exploration, individual evolution, and enforced evolution strategies. The optimized GRU network then predicts each reconstructed component, and the component-wise forecasts are aggregated to yield the final PV power output. Simulation results from several representative months indicate that the proposed approach reduces RMSE by an average of 9.02% compared to comparison model and by 43.41% relative to the baseline model, demonstrating its superior forecasting capability. Additionally, the model demonstrated scalability across varying climate conditions, confirming its applicability in real-world scenarios. Full article

Background/Objectives: Diabetic neuropathy (DN), a common complication of type 2 diabetes mellitus, manifests as peripheral nerve dysfunction with symptoms such as fatigue. Although exercise effectively reduces fatigue in neuropathy patients, precise detection methods are crucial to elucidate the role of rehabilitation. Accordingly, this study aimed to evaluate fatigue in DN patients using a multimodal approach (clinical and instrumental) and to compare the efficacy of aerobic versus resistance training on fatigue parameters. Methods: Eligible DN inpatients admitted for rehabilitation at the Neuromotor Rehabilitation Unit of the IRCCS ICS Maugeri Institute of Montescano (PV) were enrolled. Inclusion criteria included age between 65 and 85 years and confirmation via the Michigan Neuropathy Screening Instrument (anamnestic section: ≥7; clinical section: ≥2.5). Patients with confounding orthopedic, neurologic, or unstable cardiopulmonary/diabetic conditions were excluded. Overall, 36 participants were randomized into two groups: 17 underwent aerobic training (treadmill), while 19 received resistance training (elastic bands), both as supplements to a standard rehabilitation program. Assessments at baseline and post-training comprised clinical measures (Borg CR10 scale, Functional Independence Measure (FIM) total and subitems, Six-Minute Walk Test (6MWT), fasting blood glucose) and instrumental evaluations (sEMG of the tibialis anterior muscle to analyze conduction velocity intercept, slope, and changes). Results: All patients completed the protocol without dropout or adverse events. Both groups demonstrated significant improvements in FIM scores and post-exercise perceived exertion over time. Instrumental sEMG analysis confirmed a physiological fatigue trend manifested as conduction velocity reduction, yet revealed no significant differences between groups. Conclusions: Multimodal assessment provides an effective means to characterize fatigue in DN patients. Both aerobic and resistance modalities enhance functional independence and fatigue perception. Its early identification enables clinicians to tailor rehabilitation strategies to overcome exercise barriers. Full article

Knowledge graph completion via link prediction is critical for intelligent equipment maintenance systems to support accurate fault diagnosis and maintenance decision making. However, existing approaches struggle to simultaneously capture local structural dependencies and perform effective multi-hop reasoning due to limited receptive fields or inefficient path exploration mechanisms. Traditional path-based methods implicitly assume path symmetry, treating all reasoning chains equally without considering their task-specific relevance. To address this issue, we propose a Graph Attention Network (GAT)-guided semantic path reasoning framework that breaks this symmetry through attention-driven asymmetric weighting, integrating local structural encoding with global multi-hop inference. The key innovation lies in a target-guided biased path sampling strategy, which transforms GAT attention weights into probabilistic transition biases, enabling adaptive exploration of high-quality semantic paths relevant to specific prediction targets. GATs learn importance-aware local representations, which guide biased random walks to efficiently sample task-relevant reasoning paths. The sampled paths are encoded and aggregated to form global semantic context representations, which are then fused with local embeddings through a gating mechanism for final link prediction. Experimental evaluations on FB15k-237, WN18RR, and a real-world equipment maintenance knowledge graph demonstrate that the proposed method consistently outperforms state-of-the-art baselines, achieving an MRR of 0.614 on the maintenance dataset and 0.485 on WN18RR. Further analysis shows that the learned path attention weights provide interpretable asymmetric reasoning evidence, enhancing transparency for safety-critical maintenance applications. Full article

Hyperparameter optimization for cryptocurrency trading strategies encounters distinct challenges owing to continuous operation, volatility rates 3–4 times higher than equity indices, and price dynamics influenced by market sentiment. Bayesian optimization (Tree-Structured Parzen Estimator, TPE) and evolutionary algorithms (Differential Evolution, DE) are great for machine learning, but there are not many systematic comparisons for trading cryptocurrencies. This research evaluates Random Sampling, TPE, and DE through 36 factorial experiments, comprising 3 trading strategies (3, 4, and 5 hyperparameters) × 3 optimizers × 4 cryptocurrency pairs (BTC/USDT, ETH/USDT, INJ/USDT, SOL/USDT), resulting in 14,400 backtesting trials with walk-forward validation. TPE won 75% of strategy–asset pairs (9 of 12), reaching 90% of optimal performance within 13–17% of trial budgets. We find strategy-specific optimizer compatibility: mean-reversion strategies show DE underperformance independent of topology (−1% to −8%), whereas trend-following strategies show consistent DE competitiveness across assets (+13% to +37%). Most notably, for the same strategy, parameter space topology differs significantly between assets (trend following: 4.6% viable on BTC to 82% on ETH = 17.8×; mean reversion: 10.8% on ETH to 92% on SOL = 8.5×), indicating that topology results from strategy–asset interaction rather than intrinsic properties. Complete testing failures and widespread severe overfitting point to regime non-stationarity as a fundamental problem. Among the contributions are: (1) evidence shows that topological effects are dominated by optimizer–strategy compatibility (DE fails on mean-reversion strategies even in 92% viable spaces, but succeeds on trend-following strategies regardless of topology, spanning 13.6–82% viable spaces); (2) this is the first systematic Bayesian versus evolutionary comparison across 4 cryptocurrency assets; (3) parameter space topology emerges from strategy–asset interaction, varying up to 17.8-fold; and (4) single-period backtests inadequately identify parameter instability. Full article

Type 2 diabetes mellitus (T2DM) is a major global public-health problem, and physical inactivity contributes to poor disease control. In settings with limited access to health services, as in the Brazilian Amazon, interventions delivered by Community Health Workers (CHWs) within Primary Health Care (PHC) may offer a pragmatic strategy to increase physical activity (PA). We aimed to evaluate the effect of a CHW-led, theory-based intervention on PA among adults with T2DM in PHC in a cluster-randomized, community-based trial. A total of 274 participants were enrolled (intervention: n = 140, control: n = 134). CHWs in the intervention group completed a blended training (e.g., asynchronous modules, printed educational materials, and hands-on guidance). They conducted six home visits over six months to support behavior change, including increased PA. PA was measured using the International PA Questionnaire (IPAQ-LF), which assessed active commuting, walking, moderate-to-vigorous PA (MVPA), and total PA. Group-by-time effects were examined using mixed-effects zero-inflated Gamma models. No significant intervention effects were observed for the conditional mean of minutes or the probability of participation in active commuting, walking, or total PA. However, for MVPA, the zero-inflated Gamma model revealed a significant intervention effect on the probability of engaging in activity. The intervention group showed a marked reduction in the likelihood of remaining at zero minutes of MVPA (Odds Ratio = 0.08; 95% CI = 0.01–0.79; p = 0.001) compared to the control group, indicating effective behavioral activation among previously inactive participants. These findings suggest that empowering CHWs to deliver structured, theory-driven interventions within PHC can reduce inactivity among high-risk adults with T2DM in underserved communities. Full article

Dry matter intake (DMI) in dairy cows is a critical indicator of nutrient intake from feed, serving as the cornerstone of precision feeding practices, playing a critical role in improving production efficiency and enhancing the quality of dairy products. To address the high costs of traditional measurement methods and the structural complexity and large parameter counts of neural network models, this study proposes a Stacking ensemble learning model to assess DMI, with model parameters optimized using the Tuna Swarm Optimization (TSO) algorithm to enhance assessment accuracy, taking cow body weight, lying duration, lying times, rumination duration, foraging duration, walking steps, and the concentrate-to-roughage feed ratio as input variables. To further improve TSO’s search efficiency and spatial exploration, this study introduces Sine–Logistic chaotic mapping, Levy flight, and Gaussian random walk strategy to optimize the TSO algorithm, developing the improved Tuna Swarm Optimization (ITSO). ITSO-optimized Stacking model achieved superior performance in DMI assessment, with an accuracy of 95.84%, significantly outperforming SVR, RF, DT, GBR, ETR, and AdaBoost models. This study provides a robust tool for precision feeding, contributing to optimizing cow feeding strategies, improving farm efficiency, and supporting sustainable dairy farming practices. Full article

(1) Background: Pulmonary hypertension (PH) is characterized by respiratory muscle weakness, limited exercise tolerance, and reduced quality of life, but inspiratory muscle training (IMT) has emerged as a potential non-pharmacological strategy to improve functional outcomes in this population. This systematic review and meta-analysis evaluated the effects of isolated IMT on respiratory function, exercise capacity, symptom burden, and safety in adults with PH. (2) Methods: A systematic search was conducted in accordance with PRISMA guidelines. Randomized controlled trials involving adults with PH who underwent isolated IMT were included, and respiratory muscle strength, spirometric parameters, exercise capacity, dyspnea, fatigue, quality of life, and adverse events were the outcomes that were assessed. Data were pooled using meta-analytic techniques where appropriate. (3) Results: A total of 130 participants, assigned to five randomized controlled trials, met the inclusion criteria. IMT significantly improved maximal inspiratory pressure (MD = +24.01 cmH₂O), maximal expiratory pressure (MD = +23.64 cmH₂O), and six-minute walk distance (MD = +60.61 m), but no significant changes were observed in spirometric indices (FEV₁%, FVC%, and FEV₁/FVC). While several individual studies demonstrated clinically relevant improvements in six-minute walk distance, the pooled analysis did not demonstrate a statistically significant effect. IMT consistently reduced dyspnea and fatigue and improved quality-of-life domains. No serious adverse events were reported, and adherence was high. (4) Conclusions: IMT is a safe and feasible adjunct intervention in PH, providing meaningful improvements in respiratory muscle strength and symptom burden. Further large-scale trials are warranted to confirm its long-term clinical benefits. Full article

Background/Objectives: Frailty and functional decline represent major challenges for aging populations, particularly among institutionalized older adults. Preserving functional capacity is essential to maintain autonomy, mobility, and quality of life. This study aimed to compare the effects of two strength training interventions—functional electromechanical dynamometer (FEMD) training and weighted vest training—on peak concentric and eccentric force during the sit-to-stand task, as well as on functional performance and body composition in frail nursing home residents. Methods: A pilot quasi-experimental study with a non-randomized control group was conducted in 19 older adults (mean age: 86.3 ± 5.8 years). Participants were allocated to FEMD training (EG1, n = 6), weighted vest training (EG2, n = 6), or a control group (CG, n = 7). Training was performed twice weekly for eight weeks. Assessments included body composition, handgrip strength, 30 s chair stand test, 3 m walking speed, and peak concentric and eccentric force during the sit-to-stand movement. Data were analyzed using mixed-model ANOVA and complementary within-group analyses. Results: No significant group × moment interactions were observed. However, EG1 demonstrated significant within-group improvements in chair stand performance (+4.8 repetitions, p = 0.006), walking speed (+0.1 m·s⁻¹, p = 0.030), concentric peak force (+46.5%, p = 0.008), and eccentric peak force (+34%, p = 0.047). EG2 showed a smaller but significant increase in eccentric peak force (+6.1%, p = 0.019), without functional improvements. Body composition changes were modest, with EG1 showing increases in weight and BMI without concomitant fat mass gains. Conclusions: In this pilot quasi-experimental study, functional electromechanical dynamometer-based training was associated with improvements in neuromuscular performance, particularly concentric peak force. However, no significant group × moment interactions were observed, indicating that differential effects between interventions cannot be established. Functional improvements should be interpreted cautiously. The present results should therefore be considered exploratory and hypothesis-generating. These findings suggest that FEMD-based training may be a feasible and potentially beneficial functional strength training strategy for frail institutionalized older adults, which should be confirmed in adequately powered randomized controlled trials. Full article

Overfitting remains a critical challenge in data-driven financial modelling, where machine learning (ML) systems learn spurious patterns in historical prices and fail out of sample and in deployment. This paper introduces the GT-Score, a composite objective function that integrates performance, statistical significance, consistency, and downside risk to guide optimization toward more robust trading strategies. This approach directly addresses critical pitfalls in quantitative strategy development, specifically data snooping during optimization and the unreliability of statistical inference under non-normal return distributions. Using historical stock data for 50 S&P 500 companies spanning 2010–2024, we conduct an empirical evaluation that includes walk-forward validation with nine sequential time splits and a Monte Carlo study with 15 random seeds across three trading strategies. In walk-forward validation, GT-Score improves the generalization ratio (validation return divided by training return) by 98% relative to baseline objective functions. Paired statistical tests on Monte Carlo out-of-sample returns indicate statistically detectable differences between objective functions (p < 0.01 for comparisons with Sortino and Simple), with small effect sizes. These results suggest that embedding an anti-overfitting structure into the objective can improve the reliability of backtests in quantitative research. Full article