Search Results (1,108)

Accurate tomato maturity detection is critical for optimizing key agricultural operations in precision agriculture, including harvesting, grading, and quality control. Despite advances in deep learning and machine vision, reliable detection in real-world environments remains challenging due to cluttered backgrounds, dense fruit clustering, and subtle color differences between maturity stages. In response to these challenges, we present TAA-YOLOv8, an attention-enhanced detection architecture integrating a novel Tomato-Adaptive Attention (TAA) module that performs sequential channel–spatial feature refinement using an adaptive 1D convolution for channel recalibration and a balanced 5 × 5 spatial kernel for improved localization, enhancing discriminative representation while preserving computational efficiency. The framework is evaluated on three datasets representing diverse agricultural environments: a newly introduced Cross-Regional Tomato dataset collected from open-field farms in Bangladesh and greenhouse facilities in Japan, and two public benchmarks, Laboro Tomato and Tomato Plantfactory. TAA-YOLOv8m outperforms baseline YOLOv8m, achieving mAP@50–95 improvements of +9.29%, +9.00%, and +6.65% with F1-scores of 0.968, 0.976, and 0.955, respectively. It further surpasses attention-enhanced variants and RT-DETR-L, and remains competitive with YOLOv11m. Gradient-Weighted Class Activation Mapping (Grad-CAM) shows concentrated fruit-centered activations, providing transparent decision-making evidence and supporting stakeholder confidence in practical deployment within vision-based agricultural management systems. Full article

Imbalanced classification remains a fundamental challenge in machine learning, where the Area Under the ROC Curve (AUC) is widely used for threshold-independent ranking evaluation, especially in AUC maximization studies. Existing AUC maximization methods suffer from two critical limitations: they rely on single fixed kernels that fail to capture complex data structures, and they incur prohibitive computational costs due to pairwise constraint construction. To address these issues, we propose the Generalized AUC Maximization Core Vector Machine (GAM-CVM), a fast imbalanced classification framework integrating multi-kernel learning with core vector machine optimization. Multiple affinity graphs are constructed from complementary perspectives and fused via cross-diffusion into a unified kernel matrix that respects the intrinsic data manifold. This fused kernel is embedded into a generalized AUC objective with a flexible ranking margin. Given the fused kernel matrix, the optimization stage of GAM-CVM achieves asymptotic linear time complexity with respect to the number of sample pairs under a fixed approximation accuracy by reformulating the learning objective as a center-constrained minimum enclosing ball problem. Extensive experiments demonstrate that GAM-CVM achieves the best overall average ranking and significantly outperforms most competing methods while maintaining the lowest optimization-stage running time. Full article

The target price for naval equipment orders is driven by the coupling of multidimensional technical and economic factors, exhibiting typical characteristics such as high dimensionality, strong nonlinearity, multicollinearity, and small-sample fluctuations. Traditional cost estimation methods struggle to achieve high-precision fitting and interpretable decision support. To address these issues, this paper constructs an integrated prediction model that combines Boruta–Lasso two-stage feature selection, grid search-optimized CatBoost, and SHAP interpretability analysis. First, the Boruta algorithm is used for rough screening of feature significance, then Lasso regression is applied for sparse fine screening, effectively eliminating redundant features and significantly mitigating multicollinearity; grid search and five-fold repeated cross-validation are employed to optimize CatBoost hyperparameters, while 10 repeated experiments with random seeds are conducted to verify model generalization robustness. SHAP is used to quantify the marginal contribution of features, revealing nonlinear associations and statistical response transition points between core features and price. This study is based on 33 publicly available real data from main combat vessels, from which 198 modeling samples were generated through interpolation-based small-sample data augmentation. The interpolated samples were only used for data augmentation and were not considered independent empirical samples. All core conclusions were validated on the 33 original real samples, and there are no missing values in the dataset. Experimental results show that the proposed model achieved the best individual results on the test set, with a coefficient of determination of R² = 0.8949, root mean square error RMSE = 0.0554, and mean absolute error MAE = 0.0476. Across 10 repeated robustness experiments, the average results were R² = 0.8828, RMSE = 0.0586, and MAE = 0.0529, with overall performance better than comparison models such as XGBoost, random forest, and standard CatBoost. Ablation experiments validated the effectiveness of the two-stage Boruta–Lasso selection strategy in improving model accuracy and stability. SHAP attribution analysis shows that full-load displacement, number of vertical missile launch cells, number of phased array radars, and combat capability are core features highly correlated with price, all showing significant nonlinear positive correlations and clear statistical response transition points. The dataset in this study has no missing values, is entirely constructed based on publicly traceable data, and does not include confidential information such as internal shipyard costs. The findings reflect statistical associations rather than causal effects. However, the sample size and ship-type coverage are limited, so the model’s applicability is somewhat constrained, and its generalization ability needs to be further verified on larger-scale, multi-ship-type independent datasets. This model combines high prediction accuracy, strong robustness, and good interpretability, providing reliable technical support for ship equipment procurement pricing demonstration, full lifecycle cost management, and scientific procurement decision-making. Full article

Due to the complexity of recent power grids, any fault can dramatically affect the system’s quality, reliability, and stability. As a result, identifying faults becomes essential to maintaining the stability and reliability of power systems within acceptable thresholds. This article presents an innovative binary classification fault detection method in recent power grids. The proposed methodology primarily consists of two preliminary stages before the training phase: data preparation and pre-training, aimed at improving the performance of the classifier. During the data preparation phase, the synthetic minority over-sampling approach balances the raw data, and the pre-training phase identifies the optimal features and hyperparameters. A novel hybrid approach combines the Grey Wolf Optimization (GWO) and Particle Swarm Optimization (PSO) methods to optimize feature selection and adjust hyperparameters. Furthermore, four machine learning models are trained and evaluated using an actual fault dataset. In addition, several evaluation criteria and receiver operating characteristic curves are used to validate the strength and robustness of the suggested method. All experimental evaluations were performed in an Azure Machine Learning Studio (AMLS) environment. The experimental results are compared to previous studies to verify the superiority of the suggested technique. Full article

Anxiety is a multidimensional construct that influences cognitive performance in complex ways, yet its factor structure and domain-specific effects remain unclear. This study examined (1) the psychometric structure of general and specific anxiety measures, (2) their associations with cognitive performance across different domains, and (3) the predictive power of machine learning models in classifying cognitive performance based on specific anxiety in different domains. A two-stage design was employed: Stage 1 (N = 500) assessed self-reported anxiety (trait, state, generalized, social, spatial, and math anxiety) via questionnaires, while Stage 2 (N = 104) involved a set of experiments measuring cognitive performance (accuracy and reaction time) across numerical, social, spatial, and control tasks. Factor analyses revealed a correlated yet distinct structure. The model treating anxiety measures as independent factors showed the best fit among tested alternatives; however, all CFA models exhibited suboptimal absolute fit indices (TLI/CFI < 0.73). Regression analyses also demonstrated domain-specific effects: after controlling for state and generalized anxiety, trait anxiety showed small but statistically significant positive associations with performance on the social task (OR = 1.03) and spatial task (OR = 1.07). Machine learning models (Random Forest, Decision Trees, SVM) demonstrated limited predictive accuracy, with ensemble methods outperforming linear models. Prediction of reaction time in cognitive tasks, based on anxiety measures, was less powerful, suggesting that non-anxiety factors play a larger role in cognitive performance. These findings highlight the importance of distinguishing between general and domain-specific anxieties in cognitive research and demonstrate the potential of a machine learning approach in modeling anxiety–performance relationships. Full article

Moldy pear core constitutes a severe internal defect that compromises fruit quality. This study proposes a nondestructive detection method for Korla pear moldy core using Vis/NIR spectral signals, aimed at supporting post-harvest quality control and automated industrial sorting. We collected spectral signals from pears and quantified the moldy pear core area to classify samples into healthy (S = 0%), slightly moldy (0 < S ≤ 10%), and severely moldy (S > 10%) categories. We constructed a three-tier comparative framework to evaluate the progression from conventional machine learning to advanced deep learning: traditional methods using univariate selection (US) and random forest (RF) for feature extraction followed by support vector machine (SVM) classification; 1D-ResNet for direct processing of spectral signals; and two-dimensional approaches transforming signals into improved gramian angular field (IGAF) or Laplacian pyramid Markov transition field (LPMTF) images processed through deep belief network (DBN), MobileNetv3, and Vision Transformer (ViT). The LPMTF-ViT combination delivered the best performance with 98.98% test accuracy and 94.44% external validation accuracy, significantly exceeding traditional approaches and 1D-ResNet. This innovative approach delivers effective technical support for early-stage, nondestructive detection of internal fruit defects. It also establishes a scalable foundation for automated industrial inspection systems, potentially reducing post-harvest losses while ensuring premium quality control in modern fruit supply chains. Full article

Background: Small cell lung cancer (SCLC) carries a dismal prognosis with limited biomarkers for risk stratification. This study aimed to identify circulating prognostic biomarkers. Methods: We prioritized SCLC risk-associated genes using Summary-data-based Mendelian Randomization of pQTL/eQTL, differential expression, and weighted gene co-expression network analysis. Five machine learning approaches were compared to develop a diagnostic model based on ACE, AGER, and IL18R1, trained on GSE149507 and validated in GSE60052. We conducted single-cell transcriptomic analysis using public datasets (GSE150766 and GSE279570) and peripheral blood mononuclear cells (PBMCs) from our extensive-stage cohort. Finally, prioritizing the lead candidate IL18R1, we enrolled a prospective clinical cohort to assess its prognostic utility. A LASSO–Cox prognostic model incorporating plasma IL18R1 and clinical variables was internally validated (7:3 split) for progression-free survival (PFS) prediction. Results: Integrative multi-omics identified ACE, AGER, and IL18R1 as SCLC-protective genes. Elastic Net machine learning identified a two-gene predictive signature (AGER and IL18R1) with robust diagnostic accuracy. Single-cell RNA sequencing revealed the predominant downregulation of ACE, AGER, and IL18R1 in T cells and alveolar type II cells from SCLC patients. PBMC analysis further supported IL18R1 downregulation in CD8⁺ T cells, NK cells, and dendritic cells. In an independent prospective cohort (n = 300), lower plasma IL18R1 levels were independently associated with shorter PFS (HR = 0.997 per unit increase; 95% CI: 0.995–0.999; and p = 0.003), with time-dependent AUCs of 0.77–0.86. Performance in limited-stage disease was inconsistent and requires further validation. A prognostic model incorporating plasma IL18R1 and 11 clinical parameters stratified patients into distinct risk groups (HR = 5.19), showing a strong discriminative ability in extensive-stage SCLC. Conclusions: We identified ACE, AGER, and IL18R1 as protective factors against SCLC progression. Integration of plasma IL18R1 with clinical parameters provides a prognostic tool for extensive-stage SCLC. Full article

Accurate in-season rice (Oryza sativa L.) yield prediction is crucial for improved nitrogen management and climate-smart decision making, yet rigorous comparative benchmarking of machine learning (ML) models using multi-temporal UAV spectral data with independent temporal validation remains limited. This study systematically evaluated four ML algorithms (Random Forest, XGBoost, Neural Network, and Linear Regression) and two Bayesian model averaging ensembles for rice yield prediction using UAV multispectral imagery. Field experiments spanning three growing seasons (2023–2025) at Louisiana State University comprised 9–10 varieties and six nitrogen rates (0–235 kg N ha⁻¹; 576 plots). Vegetation indices and spectral bands from three growth stages were extracted as predictors. Models were compared using 300 random train–test iterations with systematic hyperparameter optimization, followed by independent validation on 2025 data. Among the individual models, XGBoost achieved the highest internal accuracy (R² = 0.87, RMSE = 0.85 t ha⁻¹), substantially outperforming Linear Regression (R² = 0.66, RMSE = 1.32 t ha⁻¹), while reduced BMA reached R² = 0.89. XGBoost demonstrated robust temporal generalization (R² = 0.62, NRMSE = 8.47%) despite environmental variation. The Enhanced Vegetation Index and Normalized Difference Red Edge at 90 days after planting (reproductive stage) were the most stable predictors across 300 iterations. Tree-based ML models substantially outperform traditional linear approaches, providing reliable pre-harvest yield forecasting for operational precision rice production. Full article

Background/Objectives: Poly(lactic-co-glycolic acid) (PLGA) microspheres are widely used in long-acting injectable (LAI) formulations because PLGA exhibits well-established biocompatibility and undergoes controlled hydrolytic degradation into metabolizable byproducts. However, optimization of microspheres typically requires time-consuming in vitro testing. Therefore, we developed a predictive machine learning model for early-stage and full time-dependent release profiles of drug-loaded PLGA microspheres. Methods: Using a published dataset comprising 321 release profiles from 89 drugs, we first developed a classification model to identify slow-release behavior (≤20% release within 3 days) and subsequently integrated the predicted early-release probability into a regression model to estimate cumulative release over time. Results: Among tree-based ensemble models, XGBoost achieved the lowest mean absolute error (MAE = 0.126) and highest Pearson correlation coefficient (r = 0.831). SHapley Additive exPlanations (SHAP) analysis revealed that drug and polymer molecular weight, predictive slow-release probability, and polymer concentration substantially influence release behavior. We also assessed this framework with external datasets. Drug release data for olaparib-loaded PLGA microspheres were obtained in-house, whereas those for semaglutide-based microspheres were obtained from the published literature. In both datasets, this framework demonstrated low MAE values (0.096 and 0.068, respectively). Conclusions: This suggests that the proposed framework can predict in vitro drug release and support efficient optimization of PLGA-based LAI formulations. Full article

Introduction: Lung cancer remains highly lethal. Endobronchial ultrasound (EBUS) enables minimally invasive diagnosis and staging. Artificial intelligence (AI) improves image analysis and diagnostic accuracy, though current evidence is limited by retrospective, small, single center studies. Methods: A scoping review following Arksey–O’Malley, Levac, and JBI frameworks, was reported as per PRISMA-ScR. Databases were searched for studies (2015–2026) on AI in EBUS. Two reviewers screened, extracted standardized data, and performed narrative synthesis grouped by algorithm type, application, and performance metrics. Results: A total of 26 studies were included. Of these, 73.1% (19/26) employed deep learning-based models, while 26.9% (7/26) used traditional or hybrid machine learning approaches. The most frequent clinical objective was diagnostic classification of malignancy (14/26; 53.8%), followed by segmentation or cytological analysis (5/26; 19.2%), anatomical navigation or lymph node station classification (3/26; 11.5%), and multimodal predictive or staging support models (4/26; 15.4%). Most studies were based on EBUS-derived images or videos (18/26; 69.2%), including both convex-probe and radial-probe applications. Studies were distributed among Convex Probe-EBUS for mediastinal staging, Radial Probe-EBUS for peripheral lesion assessment, and rapid on-site evaluation-based cytology analysis, reflecting diverse clinical contexts. Most models were developed using static images. Conclusions: AI applications in EBUS are predominantly based on deep learning and mainly focused on diagnostic classification, with growing but still limited exploration of segmentation, navigation, and multimodal approaches. The evidence reflects diverse clinical contexts and data sources, particularly image-based inputs, but remains unevenly distributed across applications. Full article

Background/Objectives: Lung cancer remains a leading cause of cancer-related death, and prognostic assessment relies mainly on TNM staging, which incompletely captures patient heterogeneity. Machine learning (ML) methods may improve survival prediction, but their use in real-world national registries with rigorous validation remains limited. This study aimed to develop ML-derived phenotypes and 1-year mortality risk groups and to evaluate their performance and clinical utility in a national lung cancer cohort from Kazakhstan. Methods: We conducted a retrospective study using a national registry including 13,685 patients. Eight routinely collected predictors were analyzed. K-means clustering was used for exploratory phenotyping. A random survival forest (RSF) model estimated 1-year mortality risk and defined low, intermediate, and high risk groups. Performance was evaluated using temporal validation, cross-validation, and bootstrap correction. Discrimination was assessed using the concordance index, prediction accuracy using the Brier score, and calibration using risk group comparisons. Comparator models included penalized Cox and TNM-only models. Clinical utility was assessed using decision-curve analysis. Results: Two phenotypes showed distinct survival outcomes, although cluster separation was modest. The RSF model showed stable performance (C-index 0.679; corrected 0.663). Risk groups demonstrated strong survival separation (high vs. low: HR 5.66). The RSF model outperformed the penalized Cox (C-index 0.544) and TNM (0.606), with improved accuracy (Brier 0.169 vs. 0.212). Calibration was generally good. Decision-curve analysis showed greater net benefit. Conclusions: An RSF-based model using routine registry data provided robust internally validated risk stratification and improved predictive performance. Clustering results were exploratory. External validation is re-quired before clinical implementation. Full article

Continuous estimation of oxygen uptake (VO₂) using wearable inertial sensors offers a practical alternative to laboratory-based metabolic testing but remains challenging due to the indirect relationship between kinematics and physiological demand. This study presents a lightweight two-stage pipeline for simultaneous heel-strike (HS) detection and VO₂ estimation using a single calf-mounted IMU. In Stage 1, an Extreme Learning Machine (ELM) + Random Forest (RF) ensemble achieves the highest HS detection F1-score (0.818) under leave-one-subject-out (LOSO) validation, outperforming a temporal convolutional network (TCN) deep learning baseline (F1 = 0.674), which exhibited higher variability across subjects. In Stage 2, kinematic and gait-derived features from 30 s windows are used to estimate normalized VO₂ via RF and ensemble regression under LOSO cross-validation across 24 participants. The RF model achieves a median R² of 0.687 using predicted HS (Pred-HS) events and 0.679 using ground-truth (GT) annotations, with the ensemble showing similar performance (median R² ≈ 0.675–0.691). No statistically significant difference was observed between GT-HS and Pred-HS conditions (p > 0.05). SHAP analysis identifies accelerometer variability (acc_std) and gyroscope-derived features as dominant predictors, with demographic variables contributing minimally. Overall, the results suggest that VO₂ estimation may be achieved using automatically detected gait events without manual annotation. The proposed pipeline is computationally efficient and indicates feasibility under controlled conditions, subject to further validation. Full article

While the demand for free-form architecture (FFA) has increased with advancements in computer-aided design (CAD) technology, the rationalization of complex surfaces into fabricable panels remains a significant challenge due to high production costs and technical complexity. Practical pain points, such as the prohibitive cost of unique molds and the inefficiency of manual data processing during design iterations, pose substantial economic risks. This study proposes an intelligent surface rationalization framework that integrates parametric design with machine learning algorithms in Autodesk^TM Dynamo Studio, a plug-in to Revit. A data-driven classification workflow was developed using four key geometric parameters—planarity, principal curvature (PC), Gaussian curvature (GC), and mean curvature (MC). Two unsupervised learning algorithms, a Gaussian mixture model and K-means clustering, were compared for their classification performance. As a result of two case studies, free-form surface classification by a Gaussian mixture model (CGMM) demonstrated flexibility in modeling complex surface data by probabilistically managing the uncertainty of the curvature distribution, and free-form surface classification by K-means clustering (CKC) was confirmed to be effective for the rapid classification of large-scale panel data. Optimizing the proportion of flat and single-curved panels through the proposed workflow contributes to deriving a reasonable balance between design intent and construction costs/constructability at the early design stage, and strengthening risk management capabilities for FFA. Full article

Early and reliable detection of rose diseases is important for automating plant monitoring and timely intervention throughout the crop lifecycle. In this context, leaf-image analysis combined with machine learning offers a practical approach for disease detection in roses. This study tests a binary classification framework that distinguishes diseased leaves using convolutional neural networks (CNNs). Three architectures were evaluated: a lightweight CNN trained from scratch as a baseline model, and two residual network models fine-tuned through transfer learning from weights pretrained on a large-scale visual recognition dataset. To assess robustness, two preprocessing strategies were also compared: a lightweight hue-based leaf isolation method that preserves full color information, and a grayscale conversion approach without masking. Experimental results obtained on a small held-out test set show strong classification performance across all evaluated models. At the same time, the findings indicate that additional validation is needed on more diverse datasets to confirm generalization under varying lighting conditions, background complexity, and plant growth stages. The results support the feasibility of CNN-based disease detection for roses and highlight its potential for integration into automated monitoring workflows. Full article

Seismic resilience assessment of high-rise buildings heavily relies on the deformation limits and fragility data of structural components, yet such data is still lacking for steel-reinforced concrete (SRC) columns, which are widely used in high-rise structures. To address this gap, this study establishes a test database consisting of 312 SRC column specimens, including 17 input parameters and three key experimental results, i.e., failure mode, yielding drift ratio θ_y, and ultimate drift ratio θ_u. Two machine learning (ML) frameworks are proposed and four ML models are trained and compared. It is found that the two-stage framework incorporating a failure mode classification shows only a slight improvement in the model performance. Thus, an end-to-end framework is recommended due to its simplicity and avoidance of error propagation, and RF and XGBoost models are adopted and tuned for θ_y and θ_u prediction for their optimal performance. Model interpretation is carried out using permutation importance (PI) and SHAP analyses to verify consistency with domain knowledge, with the key influencing factors identified as longitudinal reinforcement strength (f_yl) and axial load ratio (n_t) for deformation capacity, and shear-span ratio (λ) for failure mode classification. The performance of ML models is compared with conventional data-fitting approaches, and it is proven that ML models outperform conventional formulas, with the R² for predicting θ_y and θ_u improved by 26.5% and 32.9%, the RMSE reduced by 30.0% and 30.4%, and the MAPE reduced by 18.5% and 48.4%, respectively, thus providing a powerful data-driven tool for the seismic resilience assessment of SRC columns and expanding the fragility data of composite components. Full article