Search Results (13,633)

In geotechnical engineering, an accurate prediction is essential for the safety and effectiveness of construction projects. One example is the prediction of over/under-excavation volumes during drill and blast tunneling. Using machine learning (ML) models to predict over-excavation often results in low accuracy, especially in complex geological settings. This study explores how the pre-processing of measurement while drilling (MWD) data impacts the accuracy of ML models. In this work, a correlational analysis of the MWD data is used as the main pre-processing procedure. For each drilling event (single borehole), correlation coefficients are calculated and then supplied as inputs to the ML model. It is shown that the ML model’s accuracy improves when the correlation coefficients are used as inputs to the ML models. It is observed that datasets made from correlation coefficients help ML models to obtain higher generalization skills and robustness. The informational content of datasets after different pre-processing routines is compared, and it is shown that the correlation coefficient dataset retains information from the original MWD data. Full article

The primary objective of this study is to explore how machine learning techniques can be incorporated into the analysis of material deformation. Neural network algorithms are applied to the study of mechanical properties of wire rods subjected to cold plastic deformations. Specifically, this study explores how pre-trained neural networks with appropriate architecture can be exploited to predict apparently distinct but internally related features. Tentative predictions are made by observing only an insignificant cropped fraction of the material’s profile. The neural network models are trained and calibrated using 6400 image fractions with a resolution of $120\times 90$ pixels. Different architectures are developed with a focus on two particular aspects. Firstly, different possible architectures are compared, particularly between multi-output and multi-label convolutional neural networks (CNNs). Moreover, a hybrid model is employed, essentially a conjunction of a CNN with a multi-layer perceptron (MLP). The neural network’s input constitutes combined numerical and visual data, and its architecture primarily consists of seven dense layers and eight convolutional layers. By proper calibration and fine-tuning, observed improvements over the standard CNN models are reflected by good training and test accuracies in order to predict the material’s mechanical properties, with efficiency demonstrated by the loss function’s rapid convergence. Secondly, the role of the pre-training process is investigated. The obtained CNN-MLP model can inherit the learning from a pre-trained multi-label CNN, initially developed for distinct features such as localization and number of passes. It is demonstrated that the pre-training effectively accelerates the learning process for the target feature. Therefore, it is concluded that appropriate architecture design and pre-training are essential for applying machine learning techniques to realistic problems. Full article

Thermogravimetric analysis (TGA) is a technique used to investigate the thermal characteristics of materials by observing fluctuations in sample mass with changes in temperature. Amid the increasing worldwide focus on sustainable materials, biocomposites have become popular for their eco-friendly characteristics. Thermal stability is a crucial factor in determining the performance of biocomposites. The present research improved thermal properties by incorporating wheat straw residual filler into an epoxy resin matrix after various chemical treatments of wheat straw fibers, such as alkali (NaOH) or a combination of silane and alkali treatments. Machine learning (ML) analysis performed in WEKA 3.0 was conducted on thermal data derived from the thermogravimetric measurements of the biocomposites. This research took into account several factors, such as filler loading, single or dual chemical treatment, and temperature, to forecast the thermal-degradation behavior during combustion. Sixteen distinct regression models were used to predict the TGA curves. The K-Nearest Neighbor (KNN) classifier outperformed the other 15 models by achieving an R-squared value of 0.9999, indicating remarkable prediction skills. The strong correlation between the experimental data and the anticipated values confirmed the accuracy of the ML computations. Full article

Background: Population-specific reference genomes are essential for improving the accuracy and reliability of genomic analyses across diverse human populations. Although Vietnam ranks as the 16th most populous country in the world, with more than 86% of its population identifying as Kinh, studies specifically focusing on the Kinh Vietnamese reference genome remain scarce. Therefore, constructing a Kinh Vietnamese reference genome is valuable in the genetic research of Vietnamese. Methods: In this study, we combined PacBio long-read sequencing and Bionano optical mapping data to generate a de novo assembly of a Kinh Vietnamese genome (VHG), which was subsequently polished using multiple Kinh Vietnamese short-read whole-genome sequences (WGSs). Results: The final assembly, named VHG1.2, comprised 3.22 gigabase pairs of high-quality sequence data, demonstrating high accuracy (QV: 48), completeness (BUSCO: 92%), and continuity (295 super scaffolds, super scaffold N50: 50 Kbp). Using multiple bioinformatic tools for variant calling, we observed significant variants when the population-specific reference VHG1.2 was used compared to the standard reference genome hg38. Conclusions: Overall, our genome assembly demonstrates the advantages of a long-read hybrid sequencing approach for de novo assembly and highlights the benefit of using population-specific reference genomes in population genomic analysis. Full article

Background: Cancer is a leading cause of death worldwide, and its early detection is crucial for improving patient outcomes. This study aimed to develop and evaluate ensemble learning models, specifically stacking, for the accurate prediction of lung, breast, and cervical cancers using lifestyle and clinical data. Methods: 12 base learners were trained on datasets for lung, breast, and cervical cancer. Stacking ensemble models were then developed using these base learners. The models were evaluated for accuracy, precision, recall, F1-score, AUC-ROC, MCC, and kappa. An explainable AI technique, SHAP, was used to interpret model predictions. Results: The stacking ensemble model outperformed individual base learners across all three cancer types. On average, for three cancer datasets, it achieved 99.28% accuracy, 99.55% precision, 97.56% recall, and 98.49% F1-score. A similar high performance was observed in terms of AUC, Kappa, and MCC. The SHAP analysis revealed the most influential features for each cancer type, e.g., fatigue and alcohol consumption for lung cancer, worst concave points, mean concave points, and worst perimeter for breast cancer and Schiller test for cervical cancer. Conclusions: The stacking-based multi-cancer prediction model demonstrated superior accuracy and interpretability compared with traditional models. Combining diverse base learners with explainable AI offers predictive power and transparency in clinical applications. Key demographic and clinical features driving cancer risk were also identified. Further research should validate the model on more diverse populations and cancer types. Full article

This study evaluated the genetic parameters for growth and Vibrio parahaemolyticus (Vp_AHPND) resistance in both the introduced MK strain and the self-constructed GK strain of Penaeus vannamei, investigating the impact of genotyped female parents on trait estimates under a single-parent nested mating design. A total of 32 families from the MK strain and 44 families from the GK strain were analyzed. Fifty-four female parents from both strains were genotyped using the “Yellow Sea Chip No. 1” containing 10.0 K SNPs. In the MK strain, heritability estimates ranged from 0.439 to 0.458 for body weight (Bw) and from 0.308 to 0.489 for survival time (ST) and survival rates at 36 h (36 SR), 50% mortality (SS₅₀), and 60 h (60 SR). In the GK strain, heritability for Bw ranged from 0.724 to 0.726, while ST, 36 SR, SS₅₀, and 60 SR had heritability estimates between 0.370 and 0.593. Genetic correlations between Bw and ST were 0.601 to 0.622 in the MK strain and 0.742 to 0.744 in the GK strain. For Bw and survival rates, correlations ranged from 0.120 to 0.547 in the MK strain and from 0.426 to 0.906 in the GK strain. The genetic correlation between ST and survival rates was not significantly different from 1 (p > 0.05) in both strains. High Pearson correlations (0.853 to 0.997, p < 0.01) were observed among survival rates at different points. Predictive accuracies for Bw, ST, and survival rates using single-step genomic best linear unbiased prediction (ssGBLUP) were comparable to pedigree-based best linear unbiased prediction (pBLUP) in the MK strain, while in the GK strain, ssGBLUP improved predictive accuracies for Bw, ST, and SS₅₀ by 0.20%, 0.32%, and 0.38%, respectively. The results indicate that both growth and Vp_AHPND resistance have significant breeding potential. Although the genetic correlation between weight and resistance varies across different populations, there is a positive genetic correlation between these traits, supporting the feasibility of multi-trait selection. To enhance genetic accuracy, breeding programs should include more genotyped progeny. These findings also suggest that infection frequency and observation time influence resistance performance and breeding selection, emphasizing the need for a tailored resistance evaluation program to improve breeding efficiency and reduce costs. Full article

The migration of radioactive waste in geological environments often exhibits anomalies, such as tailing and early arrival. Fractional derivative models (FADE) can provide a good description of these phenomena. However, developing models for solute transport in unsaturated media using fractional derivatives remains an unexplored area. This study developed a variably saturated fractional derivative model combined with different release scenarios, to capture the abnormal increase observed in monitoring wells at a field site. The model can comprehensively simulate the migration of nuclides in the unsaturated zone (impermeable layer)—saturated zone system. This study fully analyzed the penetration of pollutants through the unsaturated zone (retardation stage), and finally the rapid lateral and rapid diffusion of pollutants along the preferential flow channels in the saturated zone. Comparative simulations indicate that the spatial nonlocalities effect of fractured weathered rock affects solute transport much more than the temporal memory effect. Therefore, a spatial fractional derivative model was selected to simulate the super-diffusive behavior in the preferential flow pathways. The overall fitness of the proposed model is good (R² ≈ 1), but the modeling accuracy will be lower with the increased distance from the waste source. The spatial differences between simulated and observed concentrations reflect the model’s limitations in long-distance simulations. Although the model reproduced the overall temporal variation of solute migration, it does not explain all the variability and uncertainty of the specific sites. Based on the sensitivity analysis, the fractional derivative parameters of the unsaturated zone show higher sensitivity than those of the saturated zone. Finally, the advantages and limitations of the fractional derivative model in radionuclide contamination prediction and remediation are discussed. In conclusion, the proposed FADE model coupled with unsaturated and saturated flow conditions, has significant application prospects in simulating nuclide migration in complex geological and hydrological environments. Full article

Background: Dermoscopy, a non-invasive diagnostic technique, is being increasingly used to evaluate cutaneous T-cell lymphomas such as mycosis fungoides (MF) and Sézary syndrome (SS). However, its diagnostic accuracy and role in staging remain underexplored. Objective: This study aimed to assess the dermoscopic patterns in MF and SS, correlating the findings with the disease stage and lesion type to evaluate dermoscopy’s diagnostic utility. Methods: A retrospective, monocentric analysis was conducted on patients with histologically confirmed MF or SS. Dermoscopic images were evaluated for vascular patterns, pigmentation, scaling, and keratin plugs. The statistical analysis assessed the correlations between these dermoscopic features and the TNMB staging and lesion type. A literature review was also performed to contextualize the findings, focusing on studies describing dermoscopic features in MF based on retrospective, prospective, and cross-sectional data. Results: The study included 30 patients with histologically confirmed MF or SS (19 males and 11 females; mean age: 64.5 years). The dermoscopic evaluation revealed that all the lesions were pigment-free, with vascular structures as the predominant feature. Linear vessels (40%) and serpentine vessels (13.3%) were the most frequently observed, along with dotted vessels (36.7%) and clods (10%). The vessel distribution was diffuse (40%) or perifollicular (36.7%), with a predominant red (56.7%) or orange (40%) background. Scaling was present in 76.7% of cases, either diffuse (40%) or perifollicular (36.7%), and keratin plugs were detected in 40% of the lesions. No statistically significant correlations were found between dermoscopic features and the TNMB stage or lesion type (p > 0.05). A cluster analysis identified two patient groups with differing vascular and scaling features but no clear association with disease stage. The literature review identified studies that commonly reported features in MF dermoscopy, including fine, short linear vessels and an orange-yellow background, particularly in early-stage MF. Spermatozoa-like structures have been marked as highly specific for diagnosing MF. Some studies also suggested a transition in vascular morphology from linear vessels in early disease to branched vessels and ulceration in advanced stages. Conclusions: Our results showed some vascular patterns have some potential but lack sensitivity for staging MF and SS. The terminology used and the reproducibility of our results compared to those reported in the literature showed little consistency, with none of our cases showing spermatozoa-like structures. Moreover, the same issues with the use of non-reproducible terminology were noted across the studies because it is not standardized and due to different incongruent dermoscopic patterns. More significant prospective studies with standardized descriptors and larger groups are needed to refine its diagnostic and staging utility. Full article

The Gaussian Error Linear Unit (GELU), a crucial component of the transformer model, poses a significant challenge for hardware implementation. To address this issue, this paper proposes internal symmetry piecewise approximation (ISPA) and error peak search strategy (EPSS) for high-precision and high-efficiency implementation of the GELU activation function. ISPA only approximates the positive axis of the erf in GELU and then leverages its internal symmetry to calculate the negative axis part. With ISPA, the mean square error (MSE) between the fitted result and the true value can reach $4.29\times$${10}^{-9}$ with 16 parts of the approximation segment, outperforming the regular method, which achieves $1.19\times$${10}^{-6}$ with 16 parts. Furthermore, EPSS can automatically find suitable and high-precision intervals for our piecewise approximation method. To evaluate the effectiveness of ISPA and EPSS, we conducted experiments on three different ViT models and observed negligible loss of prediction accuracy. The hardware implementation is on an XCZU9EG FPGA running at 450 MHz. Experimental results indicate that ISPA outperforms existing methods. Full article

The independent metering control system is renowned for its ability to independently regulate the flow and pressure of various actuators, achieving high efficiency and energy savings in hydraulic systems. The high-speed digital valve is known for its fast response to control signals and precise fluid control. However, challenges such as jitter in the position control of hydraulic cylinders, unknown dead zone nonlinearity, and time variance in electro-hydraulic proportional systems necessitate further investigation. To address these issues, this study initially establishes an independent metering control system with digital valves. Based on the state space equation and Lyapunov stability judgment conditions, a high-order sliding mode controller is designed. In addition, a radial basis function (RBF) neural network is constructed to approximate uncertainties arising from the modeling process, the accuracy error indicator uses Mean Absolute Error (MAE), and a finite time generalized extended state observer (GESO) is introduced to conduct online disturbance observation for the external disturbances present within the control system. Consequently, a variable structure high-order sliding mode control strategy, augmented by RBF neural networks and finite time generalized extended state observer (RBF-GESO-SMC), is proposed. Finally, simulations and experimental verification are performed, followed by a comprehensive analysis of the experimental results. Compared with the sliding mode control (SMC), the RBF-GESO-SMC diminishes the displacement-tracking control accuracy error by 63.7%. Compared with traditional Proportional-Integral-Derivative (PID) control, it reduces the displacement-tracking control accuracy error by 78.1%. The results indicate that, through the comparison with SMC and PID control, RBF-GESO-SMC exerts significant influence on the improvement of position accuracy, anti-interference ability, transient response performance, and stability. Full article

Background: Chronic liver disease (CLD) is a major global health concern, contributing significantly to morbidity and mortality. Cirrhosis and liver cancer are the leading causes of liver-related deaths, with various etiological factors, such as metabolic disorders and alcohol-related and viral hepatitis, driving its global prevalence. Non-invasive biomarkers and scoring systems have emerged as key tools for assessing liver disease severity and differentiating CLD from cirrhosis. This study evaluates biomarkers and non-invasive scores and their utility in distinguishing CLD from cirrhosis. Methods: This retrospective observational study included 250 adult patients hospitalized between January 2021 and December 2023 at Cluj County Emergency Clinical Hospital, Romania. Patients were diagnosed with either cirrhosis or CLD of viral, autoimmune, or primary biliary cholangitis (PBC) etiology. Non-invasive biomarkers, scores, and various hemogram-derived ratios were evaluated. Statistical analysis involved descriptive statistics, comparative tests, and receiver operating characteristic (ROC) curve analysis. Results: Among the 250 patients, 137 had liver cirrhosis (54.8%) and 113 had CLD without cirrhosis (45.2%). Significant differences were observed in laboratory parameters, with cirrhosis patients showing lower hemoglobin, platelet count, and albumin levels alongside higher liver enzymes and INR values. Non-invasive scores such as APRI, FIB-4, and NFS demonstrated higher values in the cirrhosis group, indicating more advanced liver damage. Hemogram-derived ratios, particularly the neutrophil-to-lymphocyte ratio (NLR), were higher in cirrhosis patients. ROC analysis revealed that the Lok index had the highest discriminatory power (AUC 0.89), followed by the King score (AUC 0.864) and the Fibrosis index (AUC 0.856), which effectively distinguished cirrhosis from CLD. Conclusions: This study underscores the utility of non-invasive biomarkers and scoring systems in differentiating CLD from cirrhosis. The Lok index, King score, and Fibrosis index demonstrated excellent diagnostic accuracy, while hemogram-derived ratios, such as NLR, offer insights into systemic inflammation associated with liver disease progression. These findings support the integration of non-invasive markers into clinical practice for improved risk stratification and management of liver diseases. Full article

This study presents an improved methodology for celestial triangulation optimization in spacecraft navigation, addressing limitations in existing approaches. While current methods like Linear Optimal Sine Triangulation (LOST) provide statistically optimal solutions for position estimation using multiple celestial body observations, their performance can be compromised by suboptimal measurement pair selection. The proposed approach, called the Improved-LOST algorithm, introduces a systematic method for evaluating and selecting optimal measurement pairs based on a Cramér–Rao Lower-Bound (CRLB) analysis. Through theoretical analysis and numerical simulations on translunar trajectories, this study demonstrates that geometric configuration significantly influences position estimation accuracy, with error variances varying by orders of magnitude depending on observation geometry. The improved algorithm outperforms conventional implementations, particularly in scenarios with challenging geometric configurations. Simulation results along a translunar trajectory using various celestial body combinations show that the systematic selection of measurement pairs based on CRLB minimization leads to enhanced estimation accuracy compared to arbitrary pair selection. The findings provide valuable insights for autonomous navigation system design and mission planning, offering a quantitative framework for assessing and optimizing celestial triangulation performance in deep space missions. Full article

Ventilation systems are susceptible to errors, external disruptions, and nonlinear dynamics. Maintaining stable operation and regulating these dynamics require an efficient control system. This study focuses on the speed control of ventilation systems using a super twisted sliding mode observer (STSMO), which provides robust and efficient state estimation for sensorless control. Traditional SM control methods are resistant to parameter fluctuations and external disturbances but are affected by chattering, which degrades performance and can cause mechanical wear. The STSMO leverages the super twisted algorithm, a second-order SM technique, to minimize chattering while ensuring finite-time convergence and high resilience. In sensorless setups, rotor speed and flux cannot be measured directly, making their accurate estimation crucial for effective ventilation drive control. The STSMO enables real-time control by providing current and voltage estimations. It delivers precise rotor flux and speed estimations across varying motor specifications and load conditions using continuous control rules and observer-based techniques. This paper outlines the mathematical formulation of the STSMO, highlighting its noise resistance, chattering reduction, and rapid convergence. Simulation and experimental findings confirm that the proposed observer enhances sensorless ventilation performance, making it ideal for industrial applications requiring reliability, cost-effectiveness, and accuracy. Full article

The use of remote sensing images to analyze the change characteristics of large-scale aquaculture areas and monitor aquaculture violations is of great significance for exploring the law of marine aquaculture and assisting the monitoring and standardization of aquaculture areas. In this study, a violation monitoring framework for marine aquaculture areas based on image recognition using an enhanced Mask R-CNN architecture incorporating a convolutional block attention module (CBAM) and soft non-maximum suppression (Soft-NMS) is proposed and applied in Sandu’ao. The results show that the modified Mask R-CNN, when compared to the most basic Mask R-CNN model, exhibits higher accuracy in identifying marine aquaculture areas. The aquaculture patterns in the Xiapu region are characterized by two peak periods of aquaculture area fluctuations, occurring in March and October. Conversely, July marks the month with the smallest aquaculture area in the region and is influenced by factors such as water temperature and aquaculture cycle. Significant changes in the aquaculture area were observed in January, March, June, August, and October, necessitating rigorous monitoring. Furthermore, monitoring and analysis of aquaculture areas have revealed that despite the reduction in illegal aquaculture acreage since 2017 due to the implementation of functional zone planning for marine aquaculture areas, illegal aquaculture activities remain prevalent in prohibited and restricted zones in Xiapu, accounting for a considerable proportion. Full article

Tongue diagnosis is a crucial method in traditional Chinese medicine (TCM) for obtaining information about a patient’s health condition. In this study, we propose a tongue image segmentation method based on deep learning and a pixel-level tongue color classification method utilizing machine learning techniques such as support vector machine (SVM) and ridge regression. These two approaches together form a comprehensive framework that spans from tongue image acquisition to segmentation and analysis. This framework provides an objective and visualized representation of pixel-wise classification and proportion distribution within tongue images, effectively assisting TCM practitioners in diagnosing tongue conditions. It mitigates the reliance on subjective observations in traditional tongue diagnosis, reducing human bias and enhancing the objectivity of TCM diagnosis. The proposed framework consists of three main components: tongue image segmentation, pixel-wise classification, and tongue color classification. In the segmentation stage, we integrate the Segment Anything Model (SAM) into the overall segmentation network. This approach not only achieves an intersection over union (IoU) score above 0.95 across three tongue image datasets but also significantly reduces the labor-intensive annotation process required for training traditional segmentation models while improving the generalization capability of the segmentation model. For pixel-wise classification, we propose a lightweight pixel classification model based on SVM, achieving a classification accuracy of 92%. In the tongue color classification stage, we introduce a ridge regression model that classifies tongue color based on the proportion of different pixel categories. Using this method, the classification accuracy reaches 91.80%. The proposed approach enables accurate and efficient tongue image segmentation, provides an intuitive visualization of tongue color distribution, and objectively analyzes and quantifies the proportion of different tongue color categories. In the future, this framework holds potential for validation and optimization in clinical practice. Full article