Search Results (61,714)

Accurate and harmonized soil information is fundamental for sustainable agricultural management, environmental assessment, and land-use planning. This study developed a national Digital Soil Mapping (DSM) framework for mainland Spain and the Balearic Islands using a harmonized database of more than 50,000 georeferenced legacy soil observations. Seven key soil properties (clay, silt, sand, pH, organic matter, available phosphorus, and exchangeable potassium) were modelled using environmental covariates derived from Sentinel-2 imagery, terrain attributes, climate, and land-use data. Three machine-learning algorithms were evaluated, with Cubist providing the best predictive performance and therefore selected for mapping all soil properties. Model validation showed the highest accuracy for soil pH (R² = 0.60), sand (R² = 0.51), and organic matter (R² = 0.50), whereas moderate performance was obtained for silt and clay. Available phosphorus and exchangeable potassium showed lower predictive accuracy, reflecting their strong dependence on management practices. The resulting maps were generated at 20 m spatial resolution and generalized to 200 m for dissemination, providing a harmonized national soil property dataset for Spain. This operational and reproducible framework supports agricultural and environmental decision-making and can be adapted for soil mapping in other regions using open-access data and open-source tools. Full article

(1) Background: Neoadjuvant chemotherapy (NAC) is important for breast cancer, but prognosis varies widely. Ki67 and apparent diffusion coefficient (ADC) reflect proliferation and cellularity, respectively. This study evaluated the prognostic value of the Ki67/ADC ratio (KA) and post-treatment ADC change (δADC) in breast cancer patients receiving NAC, and developed a survival prediction model incorporating these indicators. (2) Methods: Two cohorts of breast cancer patients treated with NAC were collected. Pre- and post-treatment breast MRI with diffusion-weighted imaging were obtained; ADC values were measured by two blinded radiologists. KA was calculated as pre-treatment Ki67 divided by pre-treatment ADC, and δADC as post-ADC minus pre-ADC. Disease-free survival (DFS) was the primary outcome. Cox regression and a predictive Cox model were used. (3) Results: A total of 419 patients were analyzed. Both KA and δADC were associated with survival. In multivariable analysis, KA remained an independent prognostic factor (HR 0.40, 95% CI 0.19–0.84, p = 0.015). High KA was associated with worse prognosis, particularly in patients without pathological complete response. The model incorporating KA showed better predictive performance than clinicopathological variables alone and effectively stratified high- vs. low-risk patients. (4) Conclusion: KA is a promising complementary biomarker for prognosis in breast cancer patients undergoing NAC. Its integration into a prognostic model improved survival risk prediction and may aid individualized post-treatment management. Full article

Tomato is a major global crop, yet foliar diseases seriously affect yield and quality. Accurate and efficient disease identification techniques are of great significance for ensuring agricultural production safety. To address the limitations of existing methods in feature extraction capability, model complexity, and dependence on large-scale labeled data, as well as the difficulty of recognizing early-stage diseases whose visual symptoms are not yet fully developed, this paper proposes SAEFormer, a lightweight and robust disease recognition model. The model integrates a Multi-scale Selective Fusion Attention Block to enhance the ability to model multi-scale semantic information in lesion areas. In addition, by leveraging a self-supervised loss derived from dense relative localization as an auxiliary regularization term, the model’s generalization ability under limited training data is notably enhanced. To optimize normalization and improve inference efficiency, the RepBN normalization strategy is further adopted, significantly reducing computational cost while maintaining model performance. Experimental results on Dataset A show that SAEFormer achieves a Top-1 accuracy of 87.86% with 24.14 M parameters and 5.35 GFLOPs, demonstrating a favorable balance between recognition accuracy and model complexity. The training curves further indicate stable convergence during model optimization. Ablation experiments validate the complementary contributions of the proposed modules. Moreover, SAEFormer achieves competitive performance in cross-dataset evaluation on Dataset B, indicating its potential adaptability to different data distributions. Overall, SAEFormer provides an efficient approach to tomato leaf disease recognition and shows potential for deployment in precision agriculture applications. Full article

Against the background of increasing styling convergence in the new energy vehicle (NEV) market, rear-end styling has gradually become a key visual interface for communicating brand identity, shaping product differentiation, and eliciting users’ Kansei cognition. However, existing Kansei design studies on automotive styling have mainly focused on whole-vehicle forms or front-face morphology, while systematic modeling methods for local rear-end styling remain limited. Under small-sample conditions, the nonlinear mapping between the Kansei semantic space and styling parameters also faces the risk of overfitting. To address these issues, this study proposes a data-driven Kansei Engineering (KE) framework integrating non-negative matrix factorization (NMF), grey relational analysis (GRA), and deep feedforward neural network (DFNN), aiming to achieve a continuous translation from Kansei need identification to parametric scheme generation for rear-end styling. First, the original seven-dimensional Kansei evaluations were aggregated into three latent Kansei dimensions through the non-negative low-rank decomposition of NMF. Second, GRA was used to screen key morphological features and reduce modeling complexity at the feature level. Third, DFNN and random forest (RF) were constructed as prediction models, and DFNN showed better average test RMSE and R² than RF. Finally, the optimal codes predicted by the DFNN were transformed into design schemes constrained by morphological coding, and their consistency in expressing the target Kansei images was verified, thereby establishing an engineering constraint-oriented and interpretable decoding pathway distinct from free-association-based Kansei design. The ablation experiment indicates that the performance advantage of the proposed framework does not arise solely from DFNN, but from the mathematical coupling among NMF-based semantic aggregation, GRA-based feature screening, and DFNN-based nonlinear mapping. This framework reformulates Kansei design as a hierarchical decomposition and modeling process, establishing a data-driven decision-support tool jointly driven by mathematical algorithms and artificial intelligence for NEV rear-end styling design. Full article

Accurate short-term significant-wave-height forecasting is important for the safe and efficient operation of coastal and offshore activities. However, forecasting across sparse buoy networks remains challenging because wave conditions exhibit complex temporal variability and spatial dependence. This study proposes an Enhanced Crossformer, a dual-tower spatiotemporal framework that combines a Crossformer branch for modelling long-range inter-station dependencies with a Cheb–LSTM branch for capturing localised graph-based interactions. The framework was evaluated using 3-hourly significant-wave-height observations from 52 National Data Buoy Centre stations along the U.S. West Coast over the period 2020–2024. Missing observations were first estimated using a graph neural network imputation model, which achieved test-set $R^2$ values ranging from 0.85 to 0.95 on artificially masked observations. Forecasting performance was then assessed for lead times of 3–24 h and compared with LSTM, Transformer, Cheb–LSTM, and Crossformer baselines. The Enhanced Crossformer achieved the lowest MAE, MSE, and MAPE values and the highest $R^2$ and correlation values at all forecast steps. Relative to the strongest baseline, it reduced MSE by up to approximately 8.7% at short lead times and maintained positive improvements through the 24 h horizon. Forecast skill was strongest under low-to-moderate sea states but decreased for high-energy wave conditions and several sheltered stations in the Southern California Bight. These results demonstrate the potential of combining global attention with graph-based local recurrence for short-term multi-buoy wave forecasting. Full article

Polygonal patterned ground (PPG) is an important remote-sensing indicator of permafrost dynamics, yet its high-resolution distribution in alpine permafrost regions remains unknown. In this study, we developed an ensemble modeling framework to map PPG in the Three-River Source Region (TRSR), northeastern Qinghai–Tibet Plateau. A total of 4200 PPG occurrence samples and multi-source environmental variables were used to train four meta models: Random Forest (RF), Support Vector Machine (SVM), Maximum Entropy (MaxEnt), and the BIOCLIM package. Model performance was evaluated using the area under the curve (AUC) and true skill statistics (TSS). An AUC-weighted ensemble model was constructed from the best-performing models. RF showed the highest accuracy, with an AUC of 0.97 and TSS of 0.85, followed by SVM and MaxEnt. The final ensemble map achieved an overall accuracy of 93.7% and a kappa coefficient of 0.91 based on validation with high-resolution satellite imagery. The mapped PPG area was 59,082 km², accounting for 25.76% of the permafrost area and 16.01% of the TRSR. PPG was mainly distributed in the Yangtze River Source Region, with limited distribution in the Yellow River Source Region and no occurrence in the Lancang River Source Region. Compared with a previous Northern Hemisphere-scale PPG distribution product, our map reduced the estimated PPG area by 21.79% and improved local spatial detail. Variable importance analysis indicated that solar radiation, elevation, freeze–thaw indices, active-layer thickness, topography, soil moisture, and ground ice jointly controlled PPG distribution. PPG mainly occurred at elevations of 4400–5000 m, on gentle slopes of 0–3°, and in areas with 30–40% ground ice content. We also found that PPG can serve as a reliable geomorphic proxy for ice-rich and thermally sensitive permafrost in alpine regions. These findings highlight the utility of PPG for permafrost dynamic assessment and associated eco-hydrological impacts in alpine permafrost regions. Full article

Leaf Area Index (LAI) is an important biophysical parameter in studies of regional and global ecosystems. However, terrain-induced distortion of surface reflectance can reduce the reliability of vegetation indices (VIs) in characterizing the canopy structure, thereby introducing uncertainties in LAI retrieval. In this study, an LAI retrieval method for mountainous areas based on the combination of terrain-corrected VIs and the random forest algorithm was proposed. Typical topographic correction models (Cosine+C, SCS+C, and Statistical–Empirical) were applied to normalize surface reflectance, and the terrain-corrected normalized difference vegetation index (NDVI) and modified soil-adjusted vegetation index (MSAVI) were constructed accordingly. Then, random forest regression was used for LAI retrieval, and the proposed method was validated through comparisons of this LAI with the field observations and original VI-based methods. The results showed that topographic correction effectively reduced the radiometric distortions induced by topography, and the NDVI_SCSC-based retrieval method performed well under various terrain conditions (with R² and RMSE of 0.927 and 0.151, respectively). In addition, to investigate the effects of different terrain factors and illumination conditions on LAI retrieval, methods based on the original and terrain-corrected VIs were compared for surfaces with different slopes and aspects. The results revealed that the terrain-corrected VIs can improve the performance of LAI retrieval in areas with terrain-induced reflectance distortion. Finally, the optimal method successfully estimated LAI in the study area. Therefore, the proposed LAI retrieval method for mountainous areas is an effective tool for extracting surface biophysical parameters, and can provide a reliable approach for regional ecological monitoring and evaluation. Full article

The intermittent nature of solar energy requires precise capacity and battery sizing for reliable microgrid design. Extending beyond traditional annual photovoltaic maps, we introduce a methodology for high-resolution capacity sizing surfaces that integrate hourly solar irradiance and battery storage estimation from averaged meteorological data, demonstrated here for Winnipeg, Canada. Leveraging pvlib Python 3.13 library, we simulate performance across all module orientations using photovoltaic modules with 22.5% efficiency and lithium-ion batteries assuming 92% round-trip efficiency, defining core metrics—Capacity ratio, Storage ratio, Excess ratio, and Battery cycling ratio, including Battery charging and discharging C-rates—while incorporating system losses. The analysis further embeds hourly unmet load and average battery state of charge. Unlike conventional PV-yield tools and single-configuration microgrid studies, the present method combines site-specific hourly meteorological data, full tilt–azimuth evaluation, battery dispatch, and explicit compliance-linked PV and storage sizing. These intuitive capacity sizing surfaces reveal how hourly modeling resolves diurnal and seasonal structure that monthly-average tools omit; a sensitivity analysis bounds the residual effect of inter-annual variability removed by climatological averaging, enhancing reliability and battery longevity under variable target standard deviation of compliance ranging from $\sigma =5/8$ to $\sigma =3$. Open-access capacity sizing surfaces for Canada-wide at 564 stations and an open-source Python library that reproduces the methodology for any hourly meteorological dataset world-wide are provided to easily size solar PV and battery systems. Full article

Against the backdrop of global carbon-neutrality targets, the circular utilization of end-of-life power batteries has become a critical pathway for securing resources and reducing environmental risks. Digital technologies are increasingly reshaping circular supply-chain management. High-performing circular suppliers are essential to the efficient operation of digital circular supply chains. However, existing research on circular supplier selection (CSS) still faces three gaps: a theoretical gap in which digitalization is neglected in circular supply chains, an information-representation gap causing evaluation distortion, and an aggregation-mechanism gap arising from the inadequate treatment of interdependencies among decision variables. To address these issues, this study constructs a decision model based on probabilistic uncertain linguistic T-spherical fuzzy sets (PULTSFSs) and Choquet-integral-based aggregated sum product assessment (CASPAS). First, this study develops a multilevel evaluation system that closely integrates digital management and control capabilities. By combining PULTSFSs with the Choquet integral (CI), a CASPAS group decision-making model that captures interaction effects among decision elements is proposed. The strategic procurement of nickel-cobalt-manganese 811 (NCM811) cathode-material precursors by CATL is used as a case study. The results show that this indicator system is better suited to digital circular supply-chain management scenarios than traditional frameworks. In the CATL case, the recycled-material supplier h₃ achieves a comprehensive score of 0.660, 5.60% higher than that of the second-ranked alternative. By capturing interdependencies among evaluation elements, the proposed model exhibits stronger ranking discriminability and robustness than traditional independent-attribute methods, thereby providing a practical foundation and an efficient tool for digital supply-chain management in the power-battery recycling industry. Full article

Electric vehicles (EVs) are crucial for mitigating greenhouse gas emissions in urban transportation. However, their integration requires efficient charging infrastructure and allocation strategies. In this paper, five heuristic algorithms were developed to allocate EVs to urban charging stations. This allocation process incorporates critical constraints, such as user preferences and charger type compatibility, while respecting station capacities governed by power output rules. The proposed methods include four initial allocation heuristics, ranging from capacity-centric and nearest-neighbor approaches to random assignments, complemented by a local search algorithm for solution refinement. To evaluate these heuristics, an optimization model minimizing station establishment and vehicle travel costs was adapted from the literature. Computational experiments were performed on both synthetic instances and real-world case studies. The results indicate that the developed heuristics, especially when enhanced by local search, deliver high-quality, near-optimal solutions within highly competitive computational times. Consequently, this study offers a scalable decision-support tool for urban planners, demonstrating how the joint optimization of infrastructure costs and user preferences can foster sustainable urban mobility and accelerate EV adoption. Ultimately, these findings offer actionable insights for scaling heterogeneous EV infrastructure, fostering urban sustainability and mitigating transport-related carbon emissions. Full article

Most studies on vehicular cybersecurity focus on classification accuracy but fail to evaluate the applicability of the models to the edge. This paper proposes a secure edge intelligence framework for V2X/IoV intrusion and misbehavior detection, called TrustEdge-V2X, which is deployment-aware. The framework combines dataset-role qualification, attack taxonomy, AI model benchmarking, feature-budget analysis, deployment ranking based on EdgeScore, offline risk-aware orchestration, external validation, robustness testing, explainable AI, repeated-run statistical analysis and ablation studies. Three datasets are assigned different experimental roles: VeReMi_NextGen is used for core V2X/VANET misbehavior detection, CICIoT2023 is used for supporting edge/IoT intrusion experiments and HCRL_CarHacking is used for external IoV/CAN validation. LightGBM outperformed all other AI models in EdgeScore (0.9383), F1-score (0.9878), MCC (0.9758), and inference latency (0.009419 ms per sample) across all eight AI models and six scenarios on VeReMi_NextGen for binary detection. In five dataset-task cases, the accuracy-best model was different from the EdgeScore-best model, which is the most important point to note: the best model in terms of accuracy is not necessarily the best model in terms of EdgeScore. Compact feature subsets were competitive, and robustness testing demonstrated an average F1 decrease of 0.1423 when tested under stress. The orchestration layer was found to be beneficial for the tasks, but it did not always perform better than the best fixed policy. As a whole, TrustEdge-V2X offers a systematic approach to the assessment and selection of vehicular cybersecurity models based on the operational and deployment conditions, not only on the classification accuracy. Full article

Machine learning has become a widely adopted approach for URL-based phishing detection, with many studies reporting F1 scores exceeding 0.95 on benchmark datasets. However, recent adversarial machine learning research has questioned the robustness of these models, suggesting that small input perturbations can severely degrade detection performance. In this study, we argue that a substantial part of this reported vulnerability stems from the way adversarial attacks are evaluated. Specifically, many existing studies assess attacks in the feature space, where feature values are modified directly without ensuring that the resulting samples correspond to valid, functional URLs. To investigate this issue, we conduct a two-stage empirical study using both a benchmark feature dataset and a dataset of real phishing URLs. Crucially, to avoid confounding the attack space with dataset differences, we additionally evaluate both feature-space and problem-space attacks on the same real-URL dataset, using an identical model and manipulable-feature budget. Our experiments reveal a striking contrast between these evaluation settings. While feature-space attacks reduce the detection rate of a Random Forest classifier on the benchmark dataset from 0.96 to 0.36, analogous manipulations performed on real URLs have almost no effect on detection performance, as the most informative signals originate from host-related attributes that are difficult for attackers to manipulate. Building on this observation, we propose a set of robust features that capture stable domain characteristics, including lexical word validity, homoglyph disguises, brand impersonation, subdomain depth, character entropy, and transport-related signals. Incorporating these features substantially improves robustness under adversarial conditions, maintaining phishing detection rates between 0.24 and 0.76 where the lexical-only baseline deteriorates to zero under a non-adaptive attacker, while also increasing the clean-data F1 score from 0.985 to 0.994. We further evaluate an adaptive attacker that explicitly targets the proposed features; although the proposed representation raises the attacker’s cost and helps under moderate attacks, host-derived features remain the only strictly attack-invariant component, so we position the proposed features as a complement to host-based signals rather than a standalone defense. Additional analyses, including model comparison, hyperparameter sensitivity analysis, feature ablation, SHAP-based interpretation, multi-seed confidence intervals, a domain-disjoint evaluation, and host-only evaluation, consistently support the proposed approach. The findings demonstrate that problem-space evaluation provides a more realistic assessment of adversarial robustness than conventional feature-space testing and show that robust feature engineering offers a practical strategy for developing phishing detection systems that remain effective under realistic adversarial conditions. Full article

Accurate national-scale mangrove mapping remains challenging because mangroves occur in narrow intertidal belts affected by tidal variation, residual clouds, water background effects, and spectral confusion with adjacent vegetation. This study evaluated five Sentinel-2 annual compositing rules for 10 m mangrove mapping in China: MAX-KNDVI, MAX-EVI, a negative-NDWI-based composite, MAX-MFI, and Median compositing. A single shared ResNet-34 U-Net model was trained using pooled patches from the five rule-specific composites and was then applied separately to each composite. Labelled sample locations, validation data, probability threshold, and post-processing settings were kept consistent to focus the comparison on the compositing rule. Using 695 labelled patch locations and 10,354 spatially independent validation points from 29 coastal regions, the Median-based output achieved the most balanced performance among the tested rules, with an overall accuracy of 90.1% and a Kappa coefficient of 0.801. Compared with GMW v3.0, HGMF_2020, and LREIS_v2_2020, the Median-based output showed higher agreement with the independent validation samples and fewer omissions in selected fragmented coastal zones. Applying the Median configuration to annual Sentinel-2 composites from 2019 to 2023 indicated an increase in mapped mangrove area in China from 22,731.76 ha to 24,631.62 ha. These results show that annual compositing-rule selection is an important source of performance variation in Sentinel-2 deep learning-based mangrove mapping and should be considered explicitly in national-scale coastal wetland monitoring. Full article

This study evaluated the effects of Myrtus communis L. extract on growth performance, growth hormone (GH), humoral immune response, oxidative stress indicators, and serum biochemical parameters in weaned Ramlıç lambs. Forty-five clinically healthy male lambs were allocated to three groups: control (n = 9), 0.5 mL/kg extract (n = 18), and 1 mL/kg extract (n = 18). The extract was administered orally once daily for 45 days. Body weight measurements and blood sampling were performed on days 0, 15, 30, and 45. Serum samples were analyzed for GH, IgG, total antioxidant status (TAS), total oxidant status (TOS), and selected biochemical parameters, whereas oxidative stress index (OSI) was calculated from TAS and TOS values. Longitudinal effects of dose, sampling day, and their interaction were evaluated using a linear mixed-effects model. Although body weight increased in all groups, no significant differences were detected among the groups (p > 0.05). A significant day × dose interaction was observed for GH (p = 0.004). IgG concentrations were significantly higher in the 1 mL/kg group on day 30 and in both extract-treated groups on day 45 than in controls. Extract administration did not cause significant changes in TAS, TOS, OSI, liver enzymes, renal function parameters, glucose, or lipid profile. In conclusion, M. communis extract was biochemically well tolerated and was associated with increased serum IgG concentrations but did not exert a marked effect on growth performance or systemic oxidant–antioxidant balance. Further studies under different physiological conditions and with longer administration periods are needed to clarify its mechanisms of action and potential use in ruminant husbandry. Full article

Background/Objectives: Chronic diseases place sustained pressure on primary care systems worldwide. While artificial intelligence (AI) shows promise for supporting chronic disease management in primary care, the factors influencing its adoption and implementation have not been comprehensively addressed within the specific context of primary care chronic disease management. The aim of the scoping review was to identify the key drivers that influence AI adoption in primary care settings for chronic disease management. Methods: This scoping review was conducted following the PRISMA ScR guidelines. A comprehensive search of seven databases was conducted between 14 April and 14 May 2025, with an updated search on 27 February 2026 using the same search strategy and eligibility criteria. Eligible studies examined the adoption or implementation of AI for chronic disease management in primary care and were published in English from 2010 to 2026. Data were charted using a standardised extraction tool and synthesised narratively. The protocol for this scoping review was registered with INPLASY. Results: Twenty-two studies were included, identifying five domains influencing AI adoption. Technical factors included data quality, interoperability, and algorithm transparency. Human factors related to clinician trust, workload, digital literacy, and preferences for hybrid human–AI care. Legal and ethical considerations centred on privacy, accountability, fairness, and independent validation. Organisational factors involved leadership, workflow integration, and resourcing, while geographical and cultural contexts shaped readiness, acceptability, and equity. Conclusions: AI adoption in chronic disease management within primary care is shaped by multi-level, context-dependent factors. The findings suggest that implementation may require coordinated strategies encompassing robust data governance and interoperability, clear regulatory frameworks, workforce capability development, workflow-aligned hybrid models with appropriate human oversight, and sustainable resourcing and reimbursement. Future research should extend beyond technical performance to evaluate acceptability, safety, workload, equity, and longer-term organisational impacts. Full article