Abstract: Thyroid hormones control crucial physiological activities, such as metabolism, oxidative stress, erythropoiesis, thermoregulation, and organ development. Hormonal imbalances may cause serious conditions like cognitive impairment, depression, and nervous system damage. Traditional diagnostic techniques, based on hormone level measurements (TSH, T3, FT4, T4, and FTI), are usually lengthy and laborious. This study uses machine learning (ML) algorithms and feature selection based on GA to improve the accuracy and efficiency of diagnosing thyroid disorders using the UCI thyroid dataset. Five ML algorithms-LR, RF, SVM, AB, and DT- were tested using two paradigms: (1) default classifiers and (2) hybrid GA-ML models- GA-RF, GA-LR, GA-SVM, GA-DT, and GA-AB. The data pre-processed included handling missing values, feature scaling, and correlation analysis. In this case, the performance metrics used for model evaluation are accuracy, F1 Score, sensitivity, specificity, precision, and Cohen’s Kappa with 80% of the dataset to train the model and the rest 20% used to test it. Among the non-hybrid models, RF achieved the highest accuracy, which was 93.93%. The hybrid GA-RF model outperformed all others, achieving a remarkable accuracy of 97.21%, along with superior metrics across all the evaluated parameters. These findings highlight the diagnostic potential of the GA-RF model in providing faster, more accurate, and reliable thyroid disorder detection. The research illustrated the potential of the hybrid GA-ML approaches to improving the clinical diagnostic process while proposing a strong and scalable approach towards thyroid disorder identification.

Abstract: Background and Objective: Lung cancer remains a leading cause of cancer-related mortality worldwide, necessitating early and accurate detection methods. Our study aims to enhance lung cancer detection by integrating VGGNet-16 form of Convolutional Neural Networks (CNNs) and Support Vector Machines (SVM) into a hybrid model (SVMVGGNet-16), leveraging the strengths of both models for high accuracy and reliability in classifying lung cancer types in different 4 classes such as adenocarcinoma (ADC), large cell carcinoma (LCC), Normal, and squamous cell carcinoma (SCC). Methods: Using the LIDC-IDRI dataset, we pre-processed images with a median filter and histogram equalization, segmented lung tumors through thresholding and edge detection, and extracted geometric features such as area, perimeter, eccentricity, compactness, and circularity. VGGNet-16 and SVM employed for feature extraction and classification, respectively. Performance matrices were evaluated using accuracy, AUC, recall, precision, and F1-score. Both VGGNet-16 and SVM underwent comparative analysis during the training, validation, and testing phases. Results: The SVMVGGNet-16 model outperformed both, with a training accuracy (97.22%), AUC (99.42%), recall (94.22%), precision (95.28%), and F1- score (94.68%). In testing, our SVMVGGNet-16 model maintained high accuracy (96.72%), with an AUC (96.87%), recall (84.67%), precision (87.40%), and F1-score (85.73%). Conclusion: Our experimental results demonstrate the potential of SVMVGGNet-16 in improving diagnostic performance, leading to earlier detection and better treatment outcomes. Future work includes refining the model, expanding datasets, conducting clinical trials, and integrating the system into clinical practice to ensure practical usability.

Abstract: In the educational landscape, educational data mining has emerged as an indispensable tool for institutions seeking to deliver exceptional and high-quality education. However, education data revealed suboptimal academic performance among a significant portion of the student population, which consequently resulted in delayed graduation. This experimental research generally aims to evaluate student graduation outcomes. Meanwhile, the specific aim is to predict student academic performance by applying the support vector machine (SVM) model based on sampling techniques. The proposed model is evaluated using datasets originating from one of the State Islamic Universities. The dataset has both on-time and delayed graduation status. The results show that the support vector machine model based on the shuffle sampling on the Arabic language and literature (BSA) dataset produces excellent performance on both tests with accuracy values above 90% and area under the curve (AUC) above 0.9. Meanwhile, the Islamic education management (MPI) dataset produces excellent performance when applying a support vector machine based on stratified sampling with accuracy values above 90% and AUC above 0.9. Therefore, it could be concluded that the proposed model has excellent and reliable performance.

Abstract: The diagnosis and treatment of epilepsy continue to face numerous challenges, highlighting the urgent need for the development of rapid, accurate, and non-invasive methods for seizure detection. In recent years, advancements in the analysis of electroencephalogram (EEG) signals have garnered widespread attention, particularly in the area of seizure recognition. A novel hybrid deep learning approach that combines feature fusion for efficient seizure detection is proposed in this study. First, the Discrete Wavelet Transform (DWT) is applied to perform a five-level decomposition of the raw EEG signals, from which time–frequency and nonlinear features are extracted from the decomposed sub-bands. To eliminate redundant features, Support Vector Machine-Recursive Feature Elimination (SVM-RFE) is employed to select the most distinctive features for fusion. Finally, seizure states are classified using Convolutional Neural Network-Bidirectional Long Short-Term Memory (CNN-Bi-LSTM). The method was rigorously validated on the Bonn and New Delhi datasets. In the binary classification tasks, both the D-E group (Bonn dataset) and the Interictal-Ictal group (New Delhi dataset) achieved 100% accuracy, 100% sensitivity, 100% specificity, 100% precision, and 100% F1-score. In the three-class classification task A-D-E on the Bonn dataset, the model performed excellently, achieving 96.19% accuracy, 95.08% sensitivity, 97.34% specificity, 97.49% precision, and 96.18% F1-score. In addition, the proposed method was further validated on the larger and more clinically relevant CHB-MIT dataset, achieving average metrics of 98.43% accuracy, 97.84% sensitivity, 99.21% specificity, 99.14% precision, and an F1 score of 98.39%. Compared to existing literature, our method outperformed several recent studies in similar classification tasks, underscoring the effectiveness and advancement of the approach presented in this research. The findings indicate that the proposed method demonstrates a high level of effectiveness in detecting seizures, which is a crucial aspect of managing epilepsy. By improving the accuracy of seizure detection, this method has the potential to significantly enhance the process of diagnosing and treating individuals affected by epilepsy. This advancement could lead to more tailored treatment plans, timely interventions, and ultimately, better quality of life for patients.

Abstract: Skin cancer is common and deadly, hence a correct diagnosis at an early age is essential. Effective therapy depends on precise classification of the several skin cancer forms, each with special traits. Because dermoscopy and other sophisticated imaging methods produce detailed lesion images, early detection has been enhanced. It's still difficult to analyze the images to differentiate benign from malignant tumors, though. Better predictive modeling methods are needed since the diagnostic procedures used now frequently produce inaccurate and inconsistent results. In dermatology, Machine learning (ML) models are becoming essential for the automatic detection and classification of skin cancer lesions from image data. With the ensemble model, which mix several ML approaches to take use of their advantages and lessen their disadvantages, this work seeks to improve skin cancer predictions. We introduce a new method, the Max Voting method, for optimization of skin cancer classification. On the HAM10000 and ISIC 2018 datasets, we trained and assessed three distinct ML models: Random Forest (RF), Multi-layer Perceptron Neural Network (MLPN), and Support Vector Machine (SVM). Overall performance was increased by the combined predictions made with the Max Voting technique. Moreover, feature vectors that were optimally produced from image data by a Genetic Algorithm (GA) were given to the ML models. We demonstrate that the Max Voting method greatly improves predictive performance, reaching an accuracy of 94.70% and producing the best results for F1-measure, recall, and precision. The most dependable and robust approach turned out to be Max Voting, which combines the benefits of numerous pre-trained ML models to provide a new and efficient method for classifying skin cancer lesions.

Abstract: Abstract Waterflooding is the most widely used improved oil recovery technique. Predicting the overall oil recovery resulting from waterflooding in oil reservoirs is crucial for effective reservoir management and appropriate decision-making. Machine learning (ML) techniques present resourceful and fast-track tools, aiding in predicting oil recovery, which is time-consuming and costly to accomplish by simulation studies. In this paper, four machine learning models: artificial neural network (ANN), Random Forest (RF), K-Nearest Neighbor (K-NN), and Support Vector Machine (SVM) are applied to estimate the overall oil recovery (R) of water flooding. Initially, statistical methods were employed to analyze the input data before applying machine learning techniques. These models take into consideration the mobility ratio (M), reservoir permeability variation (V), water-oil production ratio (WOR), and initial water saturation (S Wi ). 1054 datasets were utilized to develop machine-learning models. ANN-based correlation was developed to estimate the overall oil recovery of waterflooding. The ANN proposed model achieves a high coefficient of determination (R 2 ) of 0.999 and a low root-mean-square error (RMSE) of 0.0063 on the validation dataset. On the other hand, the other machine learning models like RF, K-NN, and SVM achieve accurate estimation of overall oil recovery (R), where the coefficients of determination (R 2 ) values are 0.97, 0.95, and 0.80 and the RMSE scores are 0.0282, 0.0405, and 0.0629 on the validation dataset, respectively. The innovative application of such ML models demonstrates significant improvements in prediction accuracy and reliability, offering a robust solution for optimizing oil recovery processes. These machine learning models provide the industry and research with efficient and economical tools for accurately estimating oil recovery in waterflooding operations within heterogeneous reservoirs.

Abstract: Abstract Evaporation represents a fundamental hydrological cycle process that demands dependable methods to quantify its fluctuation to ascertain sustainable agriculture, irrigation systems, and overall water resource management. Meteorological variables such as relative humidity, temperature, wind speed, and sunshine hours affect evaporation non-linearly, resulting in challenges while developing prediction models. To combat this, the study aimed to develop robust models for estimating evaporation in semi-arid environments by applying machine learning techniques. Daily meteorological datasets (from January 2000 to December 2010) for the above variables (input) were collected from the Sidi Yakoub meteorological station in the Wadi Sly basin, Algeria. Conventional deep neural network (DNN) coupled with support vector machine (SVM), Bayesian additive regression trees (BART), random subspace (RSS), M5 pruned, and random forest (RF) were used for developing prediction models using various input variable combinations. Model performances were compared using mean absolute error (MAE), root mean square error (RMSE), determination coefficient (R 2 ), Nash–Sutcliffe efficiency (NSE) coefficient, and percentage bias (PBIAS). Results indicated comparatively better performance for hybrid models (DNN-SVM, DNN-BART, DNN-RSS, DNN-M5 pruned, and DNN-RF) than conventional models (standalone DNN). Among hybrid models, the DNN-SVM model outperformed others with high accuracy and performance and fewer statistical errors in the daily pan evaporation prediction during the testing phase (R²=0.65, RMSE = 3.00 mm, MAE = 2.13, NSE = 0.65, and PBIAS = 3.54). DNN-RF was in the second rank for the prediction with R 2 of 0.64, RMSE of 3.00 mm, MAE of 2.16, NSE of 0.64, and PBIAS = 0.41. While the standalone DNN model gave the lowest results with MAE of 4.87, RMSE of 5.00 mm, and NRMSE of 0.65. The present framework’s success in Algeria’s Wadi Sly basin highlights its potential for scalable adoption in irrigation scheduling and drought resilience strategies, yielding implementable steps for policymakers, addressing climate-driven water scarcity. Future research should explore integrating real-time climate projections and socio-hydrological variables to improve predictive adaptability across diverse agroecological zones.

Abstract: This study presents a novel Fractional Whale Optimization Algorithm-Enhanced Support Vector Regression (FWOA-SVR) framework for solar energy forecasting, addressing the limitations of traditional SVR in modeling complex relationships within data. The proposed framework incorporates fractional calculus in the Whale Optimization Algorithm (WOA) to improve the balance between exploration and exploitation during hyperparameter tuning. The FWOA-SVR model is comprehensively evaluated against traditional SVR, Long Short-Term Memory (LSTM), and Backpropagation Neural Network (BPNN) models using training, validation, and testing datasets. Experimental results show that FWOA-SVR achieves superior performance with the lowest MSE values (0.036311, 0.03942, and 0.03825), RMSE values (0.19213, 0.19856, and 0.19577), and the highest R2 values (0.96392, 0.96104, and 0.96192) for training, validation, and testing, respectively. These results highlight the significant improvements of FWOA-SVR in prediction accuracy and efficiency, surpassing benchmark models in capturing complex patterns within the data. The findings highlight the effectiveness of integrating fractional optimization techniques into machine learning frameworks for advancing solar energy forecasting solutions.

Abstract: Knee osteoarthritis (KOA) represents a progressive degenerative disorder characterized by the gradual erosion of articular cartilage. This study aimed to develop and validate biomarker-based predictive models for KOA diagnosis using machine learning techniques. Clinical data from 2594 samples were obtained and stratified into training and validation datasets in a 7:3 ratio. Key clinical features were identified through differential analysis between KOA and control groups, combined with least absolute shrinkage and selection operator (LASSO) regression. The SHapley Additive Planning (SHAP) method was employed to rank feature importance quantitatively. Based on these rankings, predictive models were constructed using Logistic Regression (LR), Random Forest (RF), eXtreme Gradient Boosting (xGBoost), Naive Bayes (NB), Support Vector Machine (SVM), and Decision Tree (DT) algorithms. Models were developed for subsets of variables, including the top 5, top 10, top 15, and all identified features. Receiver operating characteristic (ROC) curves were applied to compare diagnostic performance across models. Additionally, a risk stratification framework for KOA prediction was designed using recursive partitioning analysis (RPA). Using difference analysis and LASSO, 44 critical clinical features were identified. Among these, age, plasma prothrombin time, gender, body mass index (BMI), and prothrombin time and international normalized ratio (PTINR) emerged as the top five features, with SHAP values of 0.1990, 0.0981, 0.0471, 0.0433, and 0.0422, respectively. Machine learning analysis demonstrated that these variables provided robust diagnostic performance for KOA. In the training set, area under the curve (AUC) values for LR, RF, xGBoost, NB, SVM, and DT models were 0.947, 0.961, 0.892, 0.952, 0.885, and 0.779, respectively. Similarly, in the validation dataset, these models achieved AUC values of 0.961, 0.943, 0.789, 0.957, 0.824, and 0.76. Among them, RF consistently exhibited superior diagnostic accuracy for KOA. Additionally, RPA analysis indicated a higher prevalence of KOA among individuals aged 54 years and older. The integration of the top five clinical variables significantly enhanced the diagnostic accuracy for KOA, particularly when employing the RF model. Moreover, the RPA model offered valuable insights to assist clinicians in refining prognostic assessments and optimizing clinical decision-making processes.

Abstract: This study considers the literature on support vector machines (SVMs) in the area of public health data analysis, particularly evaluating their ability to harness big data for disease classification and health predictions. SVMs have been remarkably embraced for two decades in clinical diagnosis, patient management, and prediction of health trends owing to their high precision and robustness. This review suggests the ability of the method to support spatially relevant health system responses through the assessment of SVM advantages and disadvantages in public health and future research agendas, including improving scalability, integrating SVMs with emerging data sources like the Internet of Things (IoT) and genomic data, and enhancing model transparency to support real-world public health decision-making.

Abstract: Over recent years, automated Human Activity Recognition (HAR) has been an area of concern for many researchers due to its widespread application in surveillance systems, healthcare environments, and many more. This has led researchers to develop coherent and robust systems that efficiently perform HAR. Although there have been many efficient systems developed to date, still, there are many issues to be addressed. There are several elements that contribute to the complexity of the task, making it more challenging to detect human activities, i.e., (i) poor lightning conditions; (ii) different viewing angles; (iii) intricate clothing styles; (iv) diverse activities with similar gestures; and (v) limited availability of large datasets. However, through effective feature extraction, we can develop resilient systems for higher accuracies. During feature extraction, we aim to extract unique key body points and full-body features that exhibit distinct attributes for each activity. Our proposed system introduces an innovative approach for the identification of human activity in outdoor and indoor settings by extracting effective spatio-temporal features, along with a Multi-Class Support Vector Machine, which enhances the model’s performance to accurately identify the activity classes. The experimental findings show that our model outperforms others in terms of classification, accuracy, and generalization, indicating its efficient analysis on benchmark datasets. Various performance metrics, including mean recognition accuracy, precision, F1 score, and recall assess the effectiveness of our model. The assessment findings show a remarkable recognition rate of around 88.61%, 87.33, 86.5%, and 81.25% on the BIT-Interaction dataset, UT-Interaction dataset, NTU RGB + D 120 dataset, and PKUMMD dataset, respectively.

Abstract: Abstract Reliable and secure wireless communication is essential in Industry 4.0. This work presents an anomaly detection framework using Digital Twin (DT) technology to simulate and monitor dynamic radio environments. By modeling network conditions and attack scenarios, the DT enables accurate identification of anomalies, particularly security threats. This study integrates machine learning with anomaly detection frameworks to enhance wireless network security. The proposed approach creates a virtual representation of the wireless environment, enabling accurate identification of anomalies and security threats. To validate the effectiveness of this framework, multiple machine learning algorithms based on traditional classifiers which are compared for their ability to detect anomalies, particularly jamming attacks. XGBoost achieved the highest accuracy (0.99) and perfect detection (1.00) of normal traffic and signal drift, outperforming Random Forest (0.98), Support Vector Machine (0.97), Logistic Regression (0.93), and K Nearest Neighbors (0.81). These results highlight XGBoost as a reliable solution for wireless network security. This work contributes to ongoing research on the integration of DT for comprehensive wireless network monitoring, emphasizing their potential to improve anomaly detection and resilience in next-generation communication systems.

Abstract: Abstract Human-induced global warming, primarily attributed to the rise in atmospheric CO 2 , poses a substantial risk to the survival of humanity. While most research focuses on predicting annual CO 2 emissions, which are crucial for setting long-term emission mitigation targets, the precise prediction of daily CO 2 emissions is equally vital for setting short-term targets. This study examines the performance of 14 models in predicting daily CO 2 emissions data from 1/1/2022 to 30/9/2023 across the top four polluting regions (China, India, the USA, and the EU27&UK). The 14 models used in the study include four statistical models (ARMA, ARIMA, SARMA, and SARIMA), three machine learning models (support vector machine (SVM), random forest (RF), and gradient boosting (GB)), and seven deep learning models (artificial neural network (ANN), recurrent neural network variations such as gated recurrent unit (GRU), long short-term memory (LSTM), bidirectional-LSTM (BILSTM), and three hybrid combinations of CNN-RNN). Performance evaluation employs four metrics ( R 2 , MAE, RMSE, and MAPE). The results show that the machine learning (ML) and deep learning (DL) models, with higher R 2 (0.714–0.932) and lower RMSE (0.480–0.247) values, respectively, outperformed the statistical model, which had R 2 (− 0.060–0.719) and RMSE (1.695–0.537) values, in predicting daily CO 2 emissions across all four regions. The performance of the ML and DL models was further enhanced by differencing, a technique that improves accuracy by ensuring stationarity and creating additional features and patterns from which the model can learn. Additionally, applying ensemble techniques such as bagging and voting improved the performance of the ML models by approximately 9.6%, whereas hybrid combinations of CNN-RNN enhanced the performance of the RNN models. In summary, the performance of both the ML and DL models was relatively similar. However, due to the high computational requirements associated with DL models, the recommended models for daily CO 2 emission prediction are ML models using the ensemble technique of voting and bagging. This model can assist in accurately forecasting daily emissions, aiding authorities in setting targets for CO 2 emission reduction.

Abstract: Knee arthritis is a prevalent joint condition that affects many people worldwide. Early detection and appropriate treatment are essential to slow the disease's progression and enhance patients’ quality of life. In this study, various machine learning and deep learning algorithms were used to detect knee arthritis. The machine learning models included k-NN, SVM, and GBM, while DenseNet, EfficientNet, and InceptionV3 were used as deep learning models. Additionally, YOLOv8 classification models (YOLOv8n-cls, YOLOv8s-cls, YOLOv8m-cls, YOLOv8l-cls, and YOLOv8x-cls) were employed. The “Annotated Dataset for Knee Arthritis Detection” with five classes (Normal, Doubtful, Mild, Moderate, Severe) and 1650 images were divided into 80% training, 10% validation, and 10% testing using the Hold-Out method. YOLOv8 models outperformed both machine learning and deep learning algorithms. k-NN, SVM, and GBM achieved success rates of 63.61%, 64.14%, and 67.36%, respectively. Among deep learning models, DenseNet, EfficientNet, and InceptionV3 achieved 62.35%, 70.59%, and 79.41%. The highest success was seen in the YOLOv8x-cls model at 86.96%, followed by YOLOv8l-cls at 86.79%, YOLOv8m-cls at 83.65%, YOLOv8s-cls at 80.37%, and YOLOv8n-cls at 77.91%.

Abstract: In credit risk assessment, unsupervised classification techniques can be introduced to reduce human resource expenses and expedite decision-making. Despite the efficacy of unsupervised learning methods in handling unlabeled datasets, their performance remains limited owing to challenges such as imbalanced data, local optima, and parameter adjustment complexities. Thus, this paper introduces a novel hybrid unsupervised classification method, named the two-stage hybrid system with spectral clustering and semi-supervised support vector machine (TSC-SVM), which effectively addresses the unsupervised imbalance problem in credit risk assessment by targeting global optimal solutions. Furthermore, a multi-view combined unsupervised method is designed to thoroughly mine data and enhance the robustness of label predictions. This method mitigates discrepancies in prediction outcomes from three distinct perspectives. The effectiveness, efficiency, and robustness of the proposed TSC-SVM model are demonstrated through various real-world applications. The proposed algorithm is anticipated to expand the customer base for financial institutions while reducing economic losses.

Abstract: Background: Hypertension is a prevalent health issue, especially among the elderly, and is linked to multiple complications. Early and accurate detection is crucial for effective management. Traditional detection methods may be limited in accuracy and efficiency, prompting the exploration of advanced computational techniques. Machine learning algorithms, combined with optimization methods, show potential in enhancing hypertension detection. Methods: In 2022, data from 1460 hypertensive and 1416 non-hypertensive individuals aged 65 and above were collected from the Lujingdong Outpatient Department of the Guangdong Second Traditional Chinese Medicine Hospital. Support Vector Machine (SVM) and Particle Swarm Optimization–Support Vector Machine (PSO-SVM) models were developed, validated using the holdout method, and evaluated based on sensitivity, specificity, positive predictive value (PPV), accuracy, G-mean, F1 score, Matthews correlation coefficient (MCC), and the area under the curve (AUC) of the receiver operating characteristic curve (ROC curve). Results: The PSO-SVM model outperformed the standard SVM, especially in sensitivity (93.9%), F1 score (0.838), and AUC-ROC (0.871). Conclusion: The PSO-SVM model is effective for complex classifications, particularly in hypertension detection, providing a basis for early diagnosis and treatment.

Abstract: Abstract The long-term performance of Continuously Reinforced Concrete Pavement (CRCP) and the optimization of maintenance strategies depend on the accurate forecasting of the International Roughness Index (IRI). For the purpose of accurately predicting the IRI in CRCP, this study offers a strong hybrid modeling framework that combines Support Vector Regression (SVR) with Genetic Algorithm (GA) optimization. Utilizing an extensive dataset from the Long-Term Pavement Performance (LTPP) program that included 395 observations and 33 CRCP sections, the suggested GA-SVR model was assessed against a number of benchmark models, such as Artificial Neural Networks (ANN), Decision Trees, Random Forests, Linear Regression, and SVR. The GA-optimized SVR model significantly outperformed all alternatives, achieving a mean RMSE of 0.039 and a coefficient of determination (R²) of 0.991 across five-fold cross-validation. Comprehensive residual analysis confirmed the model’s stability, while sensitivity analysis and feature importance rankings identified key influential variables such as Initial IRI, Layer 4 Type, and Layer 3 Thickness. Partial Dependence Plots and 3D visualizations further demonstrated how these factors affect IRI trends. The findings underscore the model’s high reliability, interpretability, and potential to support proactive pavement maintenance and design decisions. This research contributes a scalable and interpretable tool for enhancing the predictive capabilities of pavement performance models in data-driven infrastructure management.

Abstract: OBJECTIVE Given the limited accessibility of magnetic resonance imaging (MRI) for diagnosing cerebral small vessel disease (CSVD) in community settings and the lack of practical early risk assessment tools, we aimed to develop and validate an interpretable machine learning (ML) model using non-imaging routine clinical data to identify individuals at high risk of CSVD and its imaging subgroups (lacunes and white matter hyperintensities [WMH]). METHODS Data were retrospectively collected from 1,489 participants in two community-based regions of China, with CSVD diagnosed by 0.35 T mobile MRI. Data were split into development and external validation cohorts. Recursive feature elimination (RFE) was employed for predictor selection, and five ML models (Logistic Regression, Support Vector Machine, Random Forest, eXtreme Gradient Boosting, Light Gradient Boosting Machine [LightGBM]) were developed and compared using nested five-fold cross-validation. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC), accuracy, F1 score, and balanced accuracy. SHapley Additive exPlanations (SHAP) enhanced interpretability. The best-performing model was integrated with predictors selected through sequential forward selection (SFS) to build a web-based application. RESULTS The best-performing LightGBM model, evaluated across the outer folds of nested five-fold cross-validation, achieved an AUROC of 0.937 ± 0.003. In the test set, it demonstrated an AUROC of 0.862 (95 % CI 0.795-0.930). External validation in a geographically distinct region confirmed moderate generalizability with an AUROC of 0.761 (95 % CI 0.721-0.801). Among the 19 selected interpretable features, the top 10 features, age, uric acid, body mass index, creatinine, waist circumference, estimated glomerular filtration rate, fasting blood glucose, high-density lipoprotein cholesterol, homocysteine, and diastolic blood pressure, were integrated into a web-based calculator. CONCLUSIONS This study developed a reliable and interpretable LightGBM model, based on routine clinical data, offering a practical CSVD screening tool for community healthcare. Its web-based deployment provides valuable support for cerebrovascular prevention and early intervention, addressing current MRI accessibility limitations.

Abstract: Abstract Ultra-high-performance fiber-reinforced concrete (UHPFRC) is an exceptional type of cementitious composite with superior mechanical and durability performances. Achieving these properties involves maintaining a low water-to-cement ratio, optimizing aggregate size distribution, and integrating fiber reinforcement. Recently, there has been a notable trend in the development and application of UHPFRCs. However, there is still a requirement for artificial intelligence (AI) methods to predict the early-age compressive strength (CS) of UHPFRC and to define the key input factors for optimal mix design with appropriate proportions. Therefore, five AI models were chosen to assess the predictive accuracy of early-age CS in the current study. These models include support vector regression (SVR), random forest (RF), artificial neural network (ANN), gradient boosting (GB), and Gaussian Process Regression (GPR). As part of evaluating model performance and conducting error analysis, this study investigated differences in prediction accuracy among five models across training and testing datasets. Additionally, feature importance analysis was implemented to explore the influence of the input variables on the early-age CS. Results indicate that GPR and SVR models with high predictive accuracy (R 2 > 0.90) outperformed ANN, RF, and GB models. Water, superplasticizer, curing temperature, and fiber content emerged as the most significant controlling parameters affecting early-age CS. The analysis of the interaction among the significant input variables and early-age CS suggests recommended inclusion levels for optimal performance. Specifically, it is recommended that the water content be maintained between 145 and 155 kg/m 2 , the superplasticizer content between 30 and 40 kg/m 2 , and the fiber content exceed 200 kg/m 2 . These recommendations are aimed at achieving desirable early-age CS characteristics. The overall findings reveal that the AI models can effectively improve the monitoring of early-age CS of UHPFRC.

Abstract: Accurate estimation of rice leaf nitrogen content (LNC) is essential for optimizing nitrogen management in precision agriculture. However, challenges such as spectral saturation and canopy structural variations across different growth stages complicate this task. This study proposes a robust framework for LNC estimation that integrates both spectral and texture features extracted from UAV-based multispectral imagery through the development of novel Spectral–Texture Fusion Indices (STFIs). Field data were collected under nitrogen gradient treatments across three critical growth stages: heading, early filling, and late filling. A total of 18 vegetation indices (VIs), 40 texture features (TFs), and 27 STFIs were derived from UAV images. To optimize the feature set, a two-stage feature selection strategy was employed, combining Pearson correlation analysis with model-specific embedded selection methods: Recursive Feature Elimination with Cross-Validation (RFECV) for Random Forest (RF) and Extreme Gradient Boosting (XGBoost), and Sequential Forward Selection (SFS) for Support Vector Regression (SVR) and Deep Neural Networks (DNNs). The models—RFECV-RF, RFECV-XGBoost, SFS-SVR, and SFS-DNN—were evaluated using four feature configurations. The SFS-DNN model with STFIs achieved the highest prediction accuracy (R2 = 0.874, RMSE = 2.621 mg/g). SHAP analysis revealed the significant contribution of STFIs to model predictions, underscoring the effectiveness of integrating spectral and texture information. The proposed STFI-based framework demonstrates strong generalization across phenological stages and offers a scalable, interpretable approach for UAV-based nitrogen monitoring in rice production systems.