Search Results (1,158)

Individual biochemical indicators are insufficient for comprehensive tomato food flavor quality assessment, necessitating multi-parameter models of the core soluble taste matrix. We hypothesized that age stratification of trained sensory assessors would expose differential biochemical variable importance profiles in flavor quality prediction. Accordingly, this study aimed to: (1) construct and compare multiple regression models linking eight biochemical indicators to sensory scores, (2) identify key quality drivers via feature selection, and (3) examine whether age stratification alters the identified sensory drivers. Eight baseline taste indicators across 62 tomato cultivars were evaluated by 30 age-stratified trained sensory panelists (<40 and ≥40 years), using cross-validation to ensure model robustness against small-sample constraints. Partial least squares regression (PLSR), support vector regression (SVR), random forest (RF), and Boruta were applied. Random forest achieved the best performance (R² = 0.82). In the full panel model, key variables were fructose, total free amino acids, and vitamin C. After age stratification, the under-40 group retained these variables, whereas the ≥40 group replaced vitamin C with soluble solids. Fructose and total free amino acids were consistently robust drivers, while total acidity remained least important. Deploying the RF–Boruta framework within an age-stratified context provides a structured analytical framework for investigating flavor perception from biochemical data. These findings suggest that fructose and total free amino acids represent highly robust candidate indicators for flavor quality prediction, while age-stratified variances suggest the utility of integrating demographic-specific metrics into precision breeding frameworks. Full article

This study develops a leakage-free hybrid Markov–Random Forest framework for cryptocurrency price forecasting and evaluates it on Bitcoin and Ethereum. Daily OHLCV features are lagged by one trading day to prevent look-ahead bias, while regime labels are assigned from observed price changes using a two-state Markov chain with increasing and decreasing states. Regime-specific Random Forest models are then tuned independently via time-series cross-validation, allowing the predictive structure to adapt to regime-specific market conditions. The empirical results exhibit clear asset dependence. For Ethereum, the hybrid model outperforms the standalone Random Forest on magnitude-based metrics, attaining lower MAE and RMSE while also delivering a modest improvement in directional accuracy. Regime-specific tuning further identifies distinct optimal hyperparameter configurations across the increasing and decreasing states, suggesting that Ethereum’s upward and downward dynamics are structurally heterogeneous and can be better captured through regime-aware learning. By contrast, for Bitcoin, the standalone Random Forest delivers superior magnitude forecasting performance, while the regime-specific models differ only in tree depth and share the remaining tuning parameters, indicating that regime conditioning adds limited incremental value in a more persistent market. Statistical tests reinforce these findings. For Ethereum, Diebold–Mariano tests show that the hybrid significantly outperforms the standalone Random Forest under squared loss, while the absolute-loss comparison is only marginal. Across both assets, directional accuracy remains close to random chance, confirming the limited predictability of next-day price direction from lagged OHLCV features. Overall, the hybrid framework is most valuable when regime-specific dynamics are sufficiently distinct, offering improved forecasting performance and greater interpretability than a single global model. Full article

Metropolitan hinterlands are zones of intense urban–rural factor flows and spatial restructuring, where understanding rural settlement evolution is crucial for revealing human–land relationship transformations. Taking Changsha County, the core hinterland of the Changsha–Zhuzhou–Xiangtan metropolitan area in central China, as a case study, we integrated landscape pattern indices, kernel density estimation, centroid migration, the Optimal Parameters-based Geographical Detector (OPGD), and Geographically Weighted Random Forest (GWRF) to analyze the spatiotemporal evolution of rural settlements from 1990 to 2020 and identify the factors associated with the spatial differentiation of rural settlement scale in 2020. The results showed that: (1) The scale of rural settlements continuously expanded, with the total area increasing by 69.7% while patch density declined by 26.7%, exhibiting a “dense south, sparse north” pattern. High-value kernel density zones progressively clustered toward the southwestern concentric zone, and the settlement centroid persistently migrated toward the urban core. (2) The output value of secondary and tertiary industries per unit area, NDVI, and living facility adequacy were identified as the core driving factors; GDP per capita, distance to cropland, and distance to major roads also exerted notable effects, and strong synergistic interactions were detected among these factors. (3) GWRF-SHAP analysis revealed pronounced spatial heterogeneity: NDVI exhibited a south-promotion, north-suppression bidirectional effect; distance to cropland showed the most stable positive influence; road proximity was significant only at transportation hubs; the output value of secondary and tertiary industries displayed a polarized “central driving, north–south suppression” pattern; and socioeconomic factors generally stimulated expansion in suburban areas while inhibiting it in remote hinterlands. This spatial divergence can be interpreted through the “south-industry, north-agriculture” structure: suburban industrial corridors are associated with externally oriented attraction, whereas remote agricultural hinterlands are more closely related to endogenous, resource-based upgrading. The study proposes a compound explanatory framework of “natural baseline constraints–locational guidance–socioeconomic dominance,” providing a scientific basis for differentiated spatial governance of rural settlements in metropolitan hinterlands. Full article

Spatial differentiation of urban natural basement conditions leads to significant differences in urbanization development patterns and land evolution patterns in different regions. Taking Fuzhou, a typical coastal basin city located in the Minjiang River Estuary, as the study area, this paper analyzes the spatiotemporal evolution characteristics of land use/cover change (LUCC) and quantifies its driving mechanism from 1990 to 2020, by using the land use transition matrix (LUTM), the center-of-gravity model (CGM), the standard deviation ellipse (SDE), and the optimal parameters-based geographical detector (OPGD). The results show that (1) the land use structure has undergone drastic restructuring, the built-up land has increased significantly, the grassland has decreased significantly, and the cropland and forest land have shown phased evolution characteristics: a light increase from 1990 to 2000 and a continuous decline from 2000 to 2020. Water exhibited a fluctuating pattern: shrinking from 1990 to 2000, expanding from 2000 to 2010, and shrinking again from 2010 to 2020. (2) Constrained by the terrain of the Minjiang Estuary Basin, the gravity centers of cropland and grassland shifted northwestward, forest land moved southeastward, water shifted northeastward, and built-up land expanded northward. (3) Driving factors exhibited stagewise differences: socioeconomic factors played a dominant role from 1990 to 2000, with population density ($q$ = 0.4029) and nighttime light ($q$ = 0.3639) being significantly higher than other factors. From 2000 to 2010, the terrain constraint effect continued to intensify, with GDP ($q$ = 0.4470), nighttime light ($q$ = 0.3658) and DEM ($q$ = 0.3638) as the dominant factors. From 2010 to 2020, urban land pattern evolution was jointly driven by multiple factors. This study clarifies the land use evolution mechanism of coastal basin cities during urbanization, providing a scientific reference for the sustainable development of similar coastal basin cities. Full article

Fusing spatial observations from sensors with heterogeneous error characteristics is a persistent challenge in geostatistics. Classical kriging assumes a Gaussian likelihood for all observations, an assumption that fails when sensors exhibit one-sided or asymmetric noise. We present a Variable Rank Kriging (VRK) formulation that supports per-observation heterogeneous likelihoods where each observation may define its own likelihood function, thus enabling principled fusion of sensors whose noise structures are significantly different in terms of distribution family and magnitude. Within this framework, we use the exponential (one-sided) likelihood as a case study to demonstrate the method and compare it with sampling-based numerical alternatives for general likelihoods without closed forms. A non-collocated RTK calibration workflow uses kriging predictions from a sparse high-accuracy reference to characterise sensor-specific likelihood parameters without requiring co-located paired observations. Synthetic 1-D and 2-D experiments show that correct per-point likelihood specification reduces RMSE by up to 92% (1-D) and 57% (2-D) relative to a misspecified Gaussian model while also eliminating systematic positive bias. A demonstration using NEON Airborne Observation Platform lidar data at Harvard Forest confirms these findings in a practical, real-world scenario. Across multiple subsamples of the lidar dataset, the exponential likelihood reduces vegetated-zone RMSE by $20.6\%$ (open zone: $18.6\%$) and mean absolute bias by $26.5\%$ relative to a heteroscedastic Gaussian baseline. The open-source vrk Python ($>=$3.10) package provides a reproducible implementation that can be applied to any spatial domain that requires multi-sensor spatial fusion with heterogeneous error structures. Full article

Individual tree crown (ITC) segmentation in structurally complex mixed forests remains challenging under limited annotation, uneven effective height-structure support, and severe inter-crown adhesion. Existing end-to-end instance segmentation methods often require substantial instance-level annotation, and their cross-domain transferability can degrade when applied to plots with different forest structures. This study proposes a probabilistic prior-constrained instance reconstruction framework that treats semantic segmentation output as an interpretable canopy prior and reconstructs object-level crowns through a structured post-processing pipeline. A height-aware canopy support mask (HCSM) converts the probability field into a credible operational domain through hysteresis thresholding, morphological reconstruction, and a height constraint. Constrained recovery within the support domain (E2GROW) repairs coverage deficiency through spatially bounded boundary adjustment with guard rails on area ratio and buffer distance. Selective splitting then addresses residual merge errors through branch-specific seed-guided partitioning, including an aggressive Voronoi reference branch and a more conservative LOCAL/marker-controlled watershed branch with explicit trigger and child-object filtering criteria. An instance-level evaluation loop based on Gate-3 Recall, a precision proxy, and threshold-crossing audits is used during module development as an iterative safeguard. On a single 500 × 500 m mixed conifer–broadleaf plot with 306 reference crowns retained for evaluation, the high-Recall VORv1 branch improves Recall from 0.369 to 0.673 over the internal R2 baseline produced by the semantic-prior-to-instance initialization procedure, whereas the balanced E2GROW configuration achieves the highest F1_proxy with fewer predicted objects; the overall gain originates from two distinct mechanisms: threshold-crossing boundary recovery for coverage-deficient crowns and local structural decomposition for merged crown groups. Sensitivity analysis indicates that the support-domain construction is stable across the explored parameter ranges, and that the two splitting branches realize a structural Recall–precision trade-off with no evidence of simple additive gains. The framework is modular and auditable, and its demonstrated applicability is strongest for annotation-scarce closed-canopy plots where a usable semantic canopy prior and height information are available. The reported evidence represents a single-site, within-plot methodological demonstration. Full article

Water quality monitoring is a fundamental element of sustainable aquaculture management, as changes in parameters of physicochemical and biological properties directly affect the health, growth performance, and productivity of the aquaculture systems. Although traditional machine learning (ML) methods have demonstrated effectiveness in water quality classification, their performance often depends on large amounts of labeled data, which can be challenging and expensive to collect in real-world aquaculture environments. This study explores a few-shot learning (FSL) framework for data-efficient water quality classification under limited supervision to address this limitation. Several FSL models, including prototypical networks (ProtoNet), Siamese Networks, and Matching Networks were developed and evaluated in a comparative experimental framework against the traditional machine learning classifiers logistic regression, random forest, support vector machine and extreme gradient boosting. Low-data learning scenarios were simulated using a structured episodic evaluation approach. Experimental results demonstrate FSL techniques outperform traditional machine learning methods across all evaluated scenarios. Among the tested methods, ProtoNet achieved the highest performance, attaining an accuracy of 94.46% and an ROC-AUC score of 98.65%, indicating superior discriminative capability and robustness. Siamese Networks also demonstrated competitive performance under highly constrained data conditions. Furthermore, latent-space visualization, confusion matrix analysis, paired t-test statistical analysis, and ablation studies confirmed that episodic meta-learning enables the learning of highly discriminative latent representations with strong generalization capability under limited labeled data conditions. The findings highlight that FSL provides a robust and scalable framework for intelligent water quality classification in aquaculture systems, particularly in scenarios where labeled data are scarce, offering significant potential for sustainable aquaculture monitoring applications. Full article

This study presents an intelligent framework for the analysis of multidimensional geomechanical states in underground mining systems based on numerical simulation and machine learning methods. A three-dimensional geomechanical model of the Zholymbet deposit was developed in the RS3 environment using the generalized Hoek–Brown failure criterion. Numerical simulations were performed for representative mining scenarios characterized by complex excavation interaction and stress redistribution. The modelling results were transformed into a multidimensional geomechanical dataset containing stress, deformation, displacement, and yielding parameters. Principal component analysis (PCA) was applied to investigate the internal structure of the geomechanical state space and identify dominant patterns controlling the rock mass behavior. Clustering analysis revealed several geomechanical regimes corresponding to stable, transitional, and instability-prone conditions. Isolation Forest anomaly detection demonstrated that atypical geomechanical states are not randomly distributed but spatially localized near excavation systems and mining horizons. The obtained results indicate that hazardous geomechanical conditions are governed by complex interactions between stress concentration, deformation intensity, yielding processes, and excavation geometry. The proposed approach provides a basis for intelligent interpretation of large-scale numerical modelling results and may support geomechanical risk assessment in underground mining operations. Full article

Remote sensing contributes to characterizing forest structure across heterogeneous tropical regions, yet structural parameters used to compare Atlantic Forest phytophysiognomies remain limited, especially in fragmented landscapes affected by multiple drivers of forest loss and degradation. This study used Global Ecosystem Dynamics Investigation (GEDI) data to compare the structure of old-growth candidate forest polygons in four Brazilian Atlantic Forest phytophysiognomies: Dense Ombrophilous Forest (DOF), Mixed Ombrophilous Forest (MOF), Seasonal Semideciduous Forest (SSdF), and Seasonal Deciduous Forest (SDF). We analyzed canopy height (H), canopy cover (COVER), foliage height diversity (FHD), plant area index (PAI), and aboveground biomass density (AGBD) from GEDI L2B and L4A footprints acquired between 2019 and 2024. Structural differences among phytophysiognomies were significant for all variables (Kruskal–Wallis, p < 0.001), with small-to-moderate effect sizes (ε² ≈ 0.05–0.15). The strongest pairwise contrasts occurred for SDF–SSdF and SSdF–DOF, whereas MOF showed greater overlap with the other groups. Across variables, AGBD and H were the most consistent discriminators, and polygon-level summaries strengthened among-group separation. These findings show that GEDI-derived polygon-level metrics can support regional comparisons of forest structure among Atlantic Forest phytophysiognomies and help identify the strongest contrasts in fragmented landscapes. Full article

Racing pigeons (Columba livia domestica) have been selectively bred for centuries for superior flight capacity. Yet, the quantitative structure of flight performance traits and the extent to which sex influences these parameters remain poorly characterized, particularly in Turkish populations. This study aimed to evaluate flight performance in racing pigeons raised in the South Marmara region of Türkiye using three key kinematic traits (flight duration, speed, and distance) and to explore the multivariate structure and individual variation of these parameters through an integrative machine learning framework. Data were compiled from 166 individually registered pigeons (77 females, 89 males), totaling 781 race records used for pattern analysis. A composite Flight Performance Score (FPS) was constructed using min–max normalized component variables, and its internal consistency was assessed via Cronbach’s alpha and principal component analysis. Univariate comparisons revealed no statistically significant sex-related differences in any of the three flight parameters (p > 0.05 for all traits). Principal component analysis confirmed substantial overlap between male and female individuals in multivariate trait space, and Random Forest classification failed to discriminate between sexes above chance level (accuracy = 0.490; ROC-AUC = 0.500), collectively indicating that sex is not a dominant determinant of flight performance in this population. Internal consistency analysis revealed that flight duration, speed, and distance are functionally independent dimensions (Cronbach’s α = 0.135; r = −0.749 between duration and speed), with PCA of the FPS component variables indicating an effectively two-dimensional variance structure (PC1: 60.1%; PC2: 39.7%). Pattern analysis of race records identified four biologically distinct flight performance profiles, characterized by differential trade-offs among flight duration, speed, and distance, suggesting that individual-level performance strategy, rather than sex, is the primary axis of variation in this dataset. These findings challenge common breeder assumptions about sex-based differences in performance and highlight the multidimensional, individual-specific nature of flight performance in racing pigeons. Full article

The proposed methodology combines analytical control laws with adaptive mechanisms and machine-learning-assisted modules based on regression trees, random forests, and extreme gradient boosting (XGBoost). Machine learning models are employed to approximate unknown nonlinear dynamics, compensate disturbances, and adjust controller parameters, while the overall control structure is constrained by Lyapunov stability conditions. This ensures that the inclusion of data-driven components does not violate the fundamental requirement of system stability. The effectiveness of the proposed approach is evaluated through simulation experiments across three operating modes with varying degrees of nonlinearity and dynamic complexity. The results show that hybrid models incorporating ensemble machine learning methods improved performance compared with the analytical and adaptive baselines examined. XGBoost-based control achieves the lowest error values and the highest level of Lyapunov stability compliance (up to 99.3%). The main contribution of this study lies in the development of a unified synthesis framework in which machine learning is not used as a standalone control strategy but as a machine-learning-assisted support mechanism integrated into a theoretically grounded control architecture. The proposed approach provides a balance between adaptability, accuracy, and rigorous stability guarantees, suggesting potential applicability to simulation-based and offline-assisted control design tasks, while real-time embedded implementation requires additional computational optimization and validation. Full article

Estimation of rice crop yield at the village level is essential because village is the Insurance Unit (IU) for rice crop in many regions in India, and timely and accurate yield information at this scale supports timely and transparent claim settlements for farmers and supports local agricultural planning. To achieve this, a multi-source satellite data-based machine learning approach was used to estimate rice yield at the village level using optical and SAR data, climatic data and land surface model-derived parameters in Kakinada of Andhra Pradesh, India. The predictor dataset included seasonal cumulative rainfall, seasonal Normalized Difference Vegetation Index (NDVI)-Max, seasonal NDVI-Mean, seasonal Land Surface Water Index (LSWI)-Max, seasonal LSWI-Mean, season total Fraction of Absorbed Photosynthetically Active Radiation (fAPAR) and season total Root Zone Soil Moisture (RZSM), and season total backscatter of the Sentinel-1 VH polarization were used to represent crop greenness, moisture status, photosynthetic activity, soil water availability, canopy structure, and seasonal water supply. For model development and validation, village-level rice yield data from 2017 to 2023 was used, which was collected through Crop Cutting Experiment (CCE) at the maturity stage of Kharif season. In this study, four machine learning models such as Random Forest (RF), Support Vector Regression (SVR), Extreme Gradient Boosting (XGBoost), and Gradient Boosting (GB) were evaluated. The multi-source satellite data and yield data for the period 2017–2021 were used to train the models, which were independently tested on 2022 data and then applied to predict the rice yield in 2023. Leave-One-Year-Out (LOYO) cross-validation was also conducted on the 2017–2022 data to assess temporal robustness and generalization capability across years. Among the evaluated models, Random Forest exhibited the best overall performance. For the independent test year 2022, RF achieved an R² of 0.465, RMSE of 415.34 kg ha⁻¹, MAE of 322.22 kg ha⁻¹, and MAPE of 10.36%. For the prediction year 2023, RF achieved improved accuracy with an R² of 0.838, RMSE of 325.75 kg ha⁻¹, MAE of 262.21 kg ha⁻¹, and MAPE of 7.68%. Further, LOYO cross-validation also showed the robustness of RF, achieving the highest mean R² of 0.702 and mean RMSE of 384.73 kg ha⁻¹. The results illustrate that multi-source satellite data combined with machine learning can be a reliable and operationally useful tool in predicting village-level rice yield, which can be used for crop insurance claim settlement. Full article

Temperate broadleaved forests support diverse arthropod communities, but canopy-dwelling insects in European lime (Tilia cordata Mill.) stands are still poorly known. We surveyed light-attracted canopy insects in six T. cordata Genetic Conservation Units and related protected stands across Lithuania. One modified, solar-powered UV light trap was installed in the canopy (10–15 m) at each site and operated twice per month from June to August in 2023 and 2024. We used diversity metrics, similarity indices, multiple regression, and non-metric multidimensional scaling (NMDS) together with PERMANOVA to examine the structure of insect communities and assess the influence of environmental factors. In total, 6031 individuals representing 295 insect species were recorded, with higher abundance, species richness and Shannon diversity in 2024 than in 2023. Across both years and all sites, Shannon H diversity index ranged from 3.21 to 3.92. Sørensen indices indicated moderate species similarity among sites and distinct species composition at the Ukmergė genetic reserve. The 20 most abundant taxa comprised over 60% of all individuals, and dominance structure changed markedly between years: Serica brunnea dominated in 2023 but was nearly absent in 2024. Regression revealed a significant positive effect of air temperature on insect abundance (about a 31% increase per 1 °C), while precipitation had no significant effect on insect abundance. NMDS and PERMANOVA showed strong spatial structuring, with sites explaining most of the variation, and weaker but significant temporal and site-by-year effects. Overall, insect diversity metrics showed non-significant correlations with T. cordata genetic diversity parameters. Results demonstrate that mature T. cordata forest stands are important reservoirs of canopy insect diversity and highlight pronounced spatial heterogeneity, interannual dynamics, and temperature sensitivity of canopy assemblages in Lithuanian forests. Full article

Understory vegetation communities in chestnut (Castanea sativa L.) forests feature unique biodiversity patterns and high conservation value, yet the complex drivers of these communities remain poorly quantified. This study investigates the combined effects of structural, microclimatic, and topographic parameters on understory biodiversity in the mountainous region of Chalkidiki, Northern Greece. Using a nested plot design (n = 30), we integrated analytical in situ microclimatic monitoring with hemispherical photography (HemiView canopy image analysis system) to accurately quantify canopy architecture (canopy cover and solar radiation parameters), while a detailed vegetation inventory of vascular plants was performed to determine plant community structure and composition. Generalized Additive Models (GAMs) were employed to model Shannon Diversity (H’) and a weighted rarity index (RSR) representing complementary aspects of understory biodiversity. Our results reveal that the tree slenderness of the dominant stand serves as a robust proxy for stand competition and compactness. Lower slenderness values, reflecting reduced overstory competition, were significantly associated with enhanced light availability and potentially with microclimatic stability, which in turn supported higher levels of species diversity and rarity. Distinct ecological trends were observed between diversity and rarity. Shannon diversity was highest in closed forest environments characterized by lower temperatures, low stand slenderness values, southern aspects, and lower elevations, with the final model explaining 66.1% of the variance (n = 27). In contrast, species rarity was primarily driven by stand slenderness and low disturbance levels (explaining 54.6% of the variance), with the majority of rare species occurring in undisturbed stands (n = 30). These findings suggest that targeted, low-intensity management for competition promotes structurally stable stands and microclimatic buffering, facilitating the preservation of understory biodiversity. Full article

Urban canopy height is an essential indicator for characterizing vegetation structure and carbon sequestration, yet satellite LiDAR often lacks sufficient spatial resolution, airborne LiDAR is costly, and SAR has limited sensitivity to vegetation structure. This study proposes a canopy height inversion framework using high-resolution stereo pairs from the Gaofen-7 (GF-7) satellite. A 0.65 m Digital Surface Model (DSM) was generated from GF-7 data, and a relative surface height was derived by differencing the GF-7 DSM from a coarse 30 m DSM reference. Key features were selected via Boruta and Random Forest Recursive Feature Elimination (RF-RFE), and six models—linear, polynomial, support vector machine, backpropagation neural network, XGBoost, and RF—were compared. The results showed that the Boruta feature set improved average R² by 8.2%. Among all models, RF performed best (test set R² = 0.71, RMSE = 1.70 m) and exhibited the strongest resistance to overfitting. Canopy heights within Beijing’s Fifth Ring Road showed an “outer-high, inner-low” pattern: large parks exceeded 30 m, while the Central Business District remained below 3 m. GF-7 stereo pairs enable efficient and cost-effective retrieval of canopy height in fragmented urban green spaces, supporting ecological parameter quantification and urban green-space management. Full article