Search Results (5,753)

Microseismic monitoring is essential for the early warning of mine dynamic disasters; however, weak signal characteristics and strong environmental noise often lead to missed detections and false alarms. To address these challenges, this study proposes an optimized U-Net model for multi-class microseismic signal recognition under low-signal-to-noise-ratio conditions. The method combines Short-Time Fourier Transform, a U-Net encoder–decoder architecture, residual learning, and squeeze-and-excitation attention modules to enhance weak feature extraction and noise suppression. A multi-source dataset containing microseismic, knocking, blasting, noise, and earthquake signals was constructed using both field-measured data and public seismic datasets. Experimental results show that the proposed model achieved an overall validation accuracy of 99.25% and excellent recall performance for microseismic events. Under extreme noise conditions with a signal-to-noise ratio of −5 dB, the model still maintained a microseismic recognition accuracy of 98.25%. Comparative experiments further demonstrate that the integration of Short-Time Fourier Transform and residual attention modules significantly improves robustness and weak-signal discrimination capability. The proposed method provides an effective approach for intelligent microseismic monitoring and mine dynamic disaster early warning. Full article

Laser powder bed fusion (LPBF) is a widely utilized metal additive manufacturing (AM) process for fabricating intricate geometries with high mechanical strength. However, achieving defect-free parts remains challenging due to complex thermodynamics and process variability. Component quality is primarily determined by mel-pool morphology, which depends on key process parameters such as laser power, scan speed, and layer thickness. Improper parameter selection causes defects like porosity (keyhole and lack of fusion), balling, and residual stresses, compromising structural integrity. Optimizing these parameters is crucial but difficult due to the multi-scale, multi-physics nature of the process, which traditionally relies on costly, time-intensive experimental trials. We present results from a data-driven approach using machine learning (ML) models to predict and optimize LPBF melt-pool characteristics, reducing reliance on trial-and-error experimentation. We find that laser power and scan speed predominantly influence the melt-pool formation. Higher scan speeds produce more favorable melt pools, whereas excessive laser power at low scan speeds leads to deep keyhole defects. To predict and classify melt pools efficiently, several ML models are deployed, including logistic regression, decision trees, ensemble learning, and fully connected neural networks. The standard neural network achieved the highest cross-validated macro-F1 score of 0.978 ± 0.014, while the weighted neural network achieved the highest recall for the rare optimal melt-pool class, 0.967 ± 0.050. These findings show that class-weighted learning provides a recall-oriented strategy for identifying suitable LPBF process windows, while avoiding overreliance on single train-test split performance. The findings underscore the effectiveness of ML in accurately classifying LPBF melt pools to rapidly identify optimal process parameters. Full article

Background/Objectives: Clostridioides difficile infection (CDI) represents an important healthcare-associated complication in hospitalized patients, particularly in those exposed to antibiotics, prolonged hospitalization, and intensive treatment during COVID-19. This study aimed to design, evaluate, and interpret machine learning models for predicting CDI occurrence in hospitalized COVID-19 patients across individual pandemic waves, with respect to administered treatment and clinical characteristics. Methods: Anonymized clinical data from 3848 COVID-19-positive patients treated at the University Hospital of L. Pasteur in Košice, Slovakia, were analyzed following the CRISP-DM methodology. Four classification models were compared: logistic regression, Random Forest, XGBoost, and a multilayer perceptron. Missing values were addressed using MICE and KNN imputation, and class imbalance was handled through oversampling techniques. Given the low CDI prevalence of 2.68%, model performance was primarily assessed using the precision–recall area under the curve (PR-AUC), with AUROC reported for comparability. Interpretability was supported using SHAP, LIME, and odds ratio analysis. Results: The best-performing models achieved PR-AUC values up to 0.160, representing more than a fivefold improvement over the random baseline of 0.027. XGBoost reached the highest AUROC of 0.823, followed by Random Forest with 0.798. Inflammatory markers were identified as important predictors of CDI risk. A Flask-based decision-support web application was developed to provide CDI risk estimation with patient-specific explanations. A preliminary pilot usability evaluation involving two physicians yielded a mean System Usability Scale score of 73.75; however, the very small evaluator sample limits the generalizability of this finding. Conclusions: Interpretable machine learning models can support clinically meaningful CDI risk stratification in highly imbalanced COVID-19 hospital datasets. The proposed decision-support tool shows potential for future integration into clinical workflows, although external and prospective validation is required. Full article

Objectives: Secondary progressive multiple sclerosis (SPMS) is a natural transition from relapsing-remitting multiple sclerosis (RRMS) in many cases. However, whether and how these phenotypes differ on an individual basis is not fully understood, limiting timely diagnosis and management for SPMS. This study aimed to investigate how deep learning using 3-dimensional (3D) frameworks including VGG19, ResNet152, and DenseNet-121 helped differentiate SPMS from RRMS based on routine clinical datasets, and what brain areas mostly contributed to this differentiation using model explanation techniques. Methods: We examined 140 participants (70 each for RRMS and SPMS) as part of an ongoing study comprising prospectively collected clinical and imaging data from routine healthcare. The data was curated to improve consistency and completeness using different strategies and were then randomly split by subject into training (n = 120) and held-out testing (n = 20). The former was used for model development through five-fold cross validation. Deep learning used T1-weighted, T2-weighted, and FLAIR brain MRI, with optional clinical variables (n = 6). A 3D gradient-weighted class activation mapping (Grad-CAM) technique was applied to identify brain areas of significance followed by ablation studies for additional insight. Results: Among the 3D frameworks validated, VGG19 was deemed the best. Based on MRI and the best 3D VGG19 model, different data curation strategies showed largely similar results. Additionally, the models combining clinical variables with MRI achieved equivalent or slightly greater performance than MRI-only models, with an average testing area under the receiver operating characteristic curve of 0.84 when datasets were fused at the flatten layer, best at 0.92, versus 0.82 and 0.89. Model explanation indicated brain regions of significance in distinguishing SPMS from RRMS individuals, including bilateral frontal lobes, left occipital and temporal lobes, and cerebellum. Conclusions: Overall findings suggest the potential of 3D deep learning models such as VGG19 for distinguishing SPMS from RRMS using routine brain MRI and clinical data, which, along with 3D Grad-CAM, could facilitate discovery of new biomarkers underlying disease worsening. Full article

Carbon dioxide (CO₂) capture through adsorption using porous materials has emerged as a promising strategy for mitigating industrial greenhouse gas emissions. However, selecting an optimal adsorbent material under varying operating conditions remains a complex and time-consuming process when relying solely on experimental studies. In this project, a machine-learning-based framework is developed to predict CO₂ adsorption capacity and identify the most suitable adsorbent material using process and material parameters. A comprehensive dataset was constructed comprising multiple classes of adsorbent materials including activated carbon, zeolites, metal–organic frameworks (MOFs), porous organic polymers (POPs), alumina/silica, and amine-functionalized sorbents. The dataset includes key parameters such as temperature, pressure, CO₂ mole fraction, humidity, BET surface area, micropore characteristics, amine loading, heat of adsorption, particle density, pellet diameter, and bed void fraction. Two machine learning models based on the XGBoost algorithm were implemented. An XGBoost Regressor was used to predict the experimental CO₂ adsorption capacity, while an XGBoost Classifier was trained to identify the type of adsorbent used based on the input parameters. The models were trained and validated using a train–test split approach to ensure reliable performance evaluation. The results demonstrate that gradient boosting models can accurately capture complex nonlinear relationships between adsorption conditions, material properties, and adsorption performance. The developed framework provides a fast and efficient predictive tool that can assist researchers and engineers in screening adsorbent materials and optimizing CO₂ capture systems for industrial applications. Using this model, one can predict the adsorption capacity of any adsorbent used in the training dataset and predict its type with 95% accuracy. Full article

Background: Cervicogenic headache (CEH) is a secondary headache disorder originating from dysfunction in the cervical spine. In addition to pain, individuals with CEH frequently experience disturbances in postural control and sensorimotor integration, which may compromise functional capacity and quality of life. Conventional clinical assessments typically focus on pain intensity and cervical range of motion; however, these measures often fail to capture the multifactorial mechanisms underlying balance impairments in this population. Machine learning (ML) methods offer the ability to integrate multidimensional clinical data and may provide a more comprehensive approach for identifying patterns of postural stability and the factors influencing balance regulation in CEH. Methods: A secondary analysis was conducted using baseline data pooled from three registered randomized controlled trials, comprising 68 independent participants diagnosed by a neurologist according to the International Classification of Headache Disorders, 3rd edition (ICHD-3). Postural Stability Class served as the primary outcome and was derived from quantitative stability scores categorized as High, Moderate, or Low. Predictor variables included demographic characteristics (age, gender), clinical measures (pain intensity, headache frequency, symptom duration, cervical range of motion), and sensorimotor parameters (center-of-pressure sway and gaze accuracy). Five machine learning algorithms—Random Forest, XGBoost, Support Vector Machine, Logistic Regression, and Gradient Boosting—were trained and evaluated using 10-fold cross-validation with procedures implemented to reduce overfitting. Results: The Gradient Boosting classifier demonstrated the best performance, achieving an accuracy of 0.857 and an F1 score of 0.857, with a cross-validated accuracy of 0.802 ± 0.063. Random Forest and XGBoost achieved accuracies of 0.786. Feature importance analysis identified center-of-pressure sway and pain intensity as the most influential predictors of stability classification, followed by cervical flexion range of motion and gaze accuracy. Demographic variables showed minimal contribution to model performance. Conclusions: Machine learning models were able to distinguish different levels of postural stability in individuals with CEH. The findings highlight the central role of pain and sensorimotor control in balance regulation and suggest that predictive analytics may support precision physiotherapy by enabling rehabilitation strategies tailored to individual sensorimotor profiles. Full article

Background/Objective: This study evaluated the diagnostic performance of vertical growth patterns and mandibular morphology, alongside the anteroposterior dysplasia indicator (APDI), for classifying skeletal malocclusions in a Saudi adult population using cephalometric analysis. Materials and Methods: This retrospective cross-sectional discriminatory performance study analyzed 162 archived lateral cephalometric radiographs of Saudi adults aged 18–44 years. The assessed variables included Frankfort-mandibular plane angle (FMA), gonial angle, ANB angle, and APDI. Statistical analysis involved descriptive statistics, ANOVA with post hoc testing, Pearson correlation, logistic regression, and receiver operating characteristic (ROC) curve analysis. Results: Significant differences among skeletal classes were observed for all evaluated variables (p < 0.05). APDI showed the largest effect size and the highest diagnostic performance, particularly for Class III malocclusion, with excellent discriminatory ability reflected by area under the curve (AUC) values, high sensitivity, and acceptable specificity at optimal cutoff points. FMA showed moderate discriminatory performance, with higher specificity but limited sensitivity, while the gonial angle exhibited comparatively weaker diagnostic performance. In logistic regression analysis, APDI was the only significant independent associated variable of Class II malocclusion. Conclusions: Within the ANB-based classification framework used in this study, APDI showed the highest discriminatory performance for skeletal malocclusion classification, supporting its role as a primary sagittal indicator. FMA contributed adjunctive information on vertical skeletal pattern, while the gonial angle showed limited diagnostic value. Combined assessment of sagittal and vertical parameters may improve cephalometric diagnosis. Full article

Background: Gorlin–Goltz syndrome is a rare autosomal dominant condition with characteristic craniofacial and odontogenic anomalies. Orofacial alterations in childhood may precede dermatological findings, highlighting the relevance of early orthodontic and functional evaluation. Objective: This case describes a multidisciplinary orthodontic and Home Sleep Apnea Testing (HSAT)-based approach for the assessment of craniofacial morphology and sleep-disordered breathing (SDB) risk in a pediatric patient with Gorlin–Goltz syndrome. Methods: A 12-year-old male with a genetically confirmed PTCH1 mutation underwent digital intraoral scanning, orthodontic evaluation, and SDB screening using the Pediatric Sleep Questionnaire (PSQ). Following a positive screening score, HSAT with the Philips Alice NightOne^® system was performed under specialist supervision. Results: The patient showed recurrent odontogenic cysts, a lateral open bite, and unilateral Class II canine relationship. The PSQ score was 0.579, exceeding the validated cut-off of 0.33 and indicating an elevated SDB risk. HSAT findings were suggestive of mild obstructive sleep apnea based on Respiratory Event Index (REI) values (REI 4.7/h), with an isolated SpO₂ nadir of 77% and a maximum recorded apnea duration of 425 s, warranting cautious specialist interpretation and follow-up assessment. Conclusions: Integrating orthodontic assessment, digital documentation, validated screening tools, and objective HSAT-based evaluation may support the early recognition of functional compromise in syndromic pediatric patients. Positive screening results should prompt specialist referral and objective sleep assessment, while attended polysomnography remains indicated when comprehensive sleep architecture evaluation or definitive characterization is required. Full article

This study introduces a Caputo fractional-order version of the SEIRS-V model to investigate the spreading dynamics of worms within wireless sensor networks. Traditional integer-order worm propagation models describe the instantaneous evolution of network states; however, they do not adequately account for memory and hereditary characteristics that may influence the transmission dynamics. Consequently, their ability to represent realistic network behavior can be limited in systems where past states affect current propagation patterns. The framework divides sensor nodes into susceptible, exposed, infectious, recovered, and vaccinated classes, while explicitly incorporating worm transmission rates, temporary loss of immunity, and the impact of preventive security measures under limited resource conditions. A detailed theoretical examination is performed, covering the existence, boundedness, and uniqueness of solutions of the fractional-order system. The coupled nonlinear fractional system is solved semi-analytically by means of the Fractional Reduced Differential Transform (FRDT) technique. To confirm accuracy and robustness, the identical system is also discretized and solved using the finite difference scheme (FDS). Unlike previous studies on worm propagation models in wireless sensor networks, which are mainly limited to equilibrium point analysis and qualitative investigations without deriving explicit solutions, the present work develops an approximate semi-analytical solution for the fractional-order SEIRS-V system using the FRDTM. Comparisons between the two solution sets demonstrate excellent agreement and high precision. Numerical outcomes are presented through a series of 2D graphical profiles that illustrate the time-dependent behavior of each compartment and reveal the sensitivity of worm propagation and suppression to variations in the fractional order and key model parameters. The integrated theoretical and computational findings underscore the strong protective role of vaccination in mitigating worm outbreaks and offer valuable guidelines for strengthening cybersecurity measures in wireless sensor networks. Full article

Rolling bearings, as core components of rotating machinery, are prone to failure under harsh working conditions, and their fault diagnosis is crucial for the safe operation of industrial systems. Aiming at resolving the problems of weak fault feature representation, poor model generalization ability and high dependence on manual preprocessing in traditional bearing fault diagnosis methods, an end-to-end fault diagnosis model named KanMSConv is proposed for one-dimensional raw vibration signals. The model abandons complex time–frequency transformation and manual feature engineering, and constructs a multi-scale feature extraction module based on depthwise separable convolution to capture local impulsive components and global modulation characteristics of fault signals simultaneously. The SE channel attention mechanism is integrated to adaptively enhance fault-related critical features and reduce redundant channel responses. Residual connection is introduced to alleviate the gradient degradation problem of deep networks and improve feature reuse capability. On this basis, the Kolmogorov–Arnold Network (KAN) is used to replace the traditional fully connected layer, which enhances the model’s ability to fit complex nonlinear mapping relationships and distinguish fault classification boundaries. Experimental verification is carried out on three representative rolling bearing datasets (CWRU, PU, SDUST) under multi-load, multi-class and cross-platform conditions. The results show that the KanMSConv model achieves 100% accuracy on the CWRU dataset, 99.93% on the PU dataset and 99.80% on the SDUST dataset, which is significantly superior to the existing mainstream fault diagnosis models in terms of Accuracy, Precision, Recall and F1-Score. And the ablation and computational cost analyses further support this conclusion. Full article

Background: Spinocerebellar ataxia type 2 (SCA2) is an inherited neurodegenerative disorder characterized by progressive cerebellar degeneration. One difficulty in treating this disease lies in identifying preclinical carriers: individuals who carry the pathogenic ATXN2 mutation but remain asymptomatic with respect to motor manifestations. Though magnetic resonance imaging (MRI) has proven valuable in supporting the diagnosis of ataxia, traditional univariate approaches using linear measurements have shown limited ability to capture the complex anatomical changes that occur across the disease spectrum, particularly during the preclinical phase. Methods: This study employed a comprehensive multivariate approach to improve the classification of individuals across the SCA2 spectrum. We developed a multinomial logistic regression model incorporating multiple linear measurements derived from magnetic resonance imaging to discriminate between healthy controls (n = 72), preclinical carriers (n = 17), and patients with manifest SCA2 (n = 61). To mitigate inherent class imbalance, particularly in the smaller preclinical subgroup, we implemented the Synthetic Minority Over-sampling Technique (SMOTE), generating a balanced dataset that enhances the model’s ability to discern the distinctive anatomical features. This was compared to the model applied to the unbalanced data. An improvement was observed when applying SMOTE. Results: The multivariate model demonstrated discriminatory performance, achieving an overall accuracy of 80.7%. The ability to identify healthy controls (AUC: 0.96), preclinical individuals (AUC: 0.75), and clinical individuals (AUC: 95%). This represents an advance over previous univariate approaches, which have had difficulty capturing the neurodegenerative changes characteristic of the preclinical stage. Conclusions: By integrating multiple neuroimaging biomarkers into a multivariable model, this study provides a tool for early identification of preclinical SCA2 carriers. The ability to accurately classify these individuals opens an opportunity for early therapeutic intervention before irreversible neurological deterioration occurs. This approach shows promise for optimizing clinical trial design and personalized care in SCA2. Full article

The internal quality of watermelons has garnered extensive attention. Conventional destructive quality detection for watermelons causes fruit loss, while existing acoustic techniques often rely on a single evaluation index. To address these issues, this study proposes a non-destructive method for comprehensive watermelon quality detection using acoustic and vibration signals. Signals from two watermelon varieties were collected under impact excitation to extract six time-domain and frequency-domain features. Factor Analysis of Mixed Data (FAMD) was employed to integrate ripeness, Soluble Solids Content (SSC), firmness, and sensory scores into a Comprehensive Quality Index (CQI), categorizing samples into High-Quality, Medium-Quality, and Low-Quality groups. Following physically constrained data augmentation to mitigate small sample size and class imbalance, an Extremely Randomized Trees (Extra-Trees) model was constructed. Results demonstrate that the Extra-Trees model achieved an overall testing accuracy of 0.92, with recall rates of 0.93 and 1.00 for Low-Quality and High-Quality watermelons, respectively. Recognition for Medium-Quality samples was lower due to overlapping physical and acoustic characteristics. Ultimately, this system aligns with actual consumer demands, providing technical support for low-cost, portable, and non-destructive watermelon grading. Full article

The prediction of student engagement characteristics involves forecasting and analyzing student interaction with educational materials using engagement prediction models. This process encompasses the prediction of cognitive, behavioral, and emotional dimensions of engagement. The existing student engagement prediction models have some limitations, including poor convergence, less generalizability, complexity issues, overfitting, false errors, and limited resources. Hence, the research proposes the Multi-Instance Localization-based Gradient Boosted Long Short-Term Memory (MIL-GBLTM) model to tackle the challenge of predicting student engagement characteristics in online classes. The integration of effective MIL with a Triplet Attention mechanism focuses on the significant features that help with engagement prediction; LSTM layers capture intricate sequential patterns, and fractional gradient boosting is used for fine-tuning for accurate prediction, alongside ensemble-based learning. The LSTM layers with the Triplet Attention module refine temporal attention, and Fractional Gradient Boosting ensures the model’s adaptability and robustness. By combining these components, the proposed model is able to predict accurate student engagement with high convergence. This integrated approach enhances the capabilities of engagement prediction models in educational contexts, facilitating more effective interventions and personalized student support in online learning environments. Experimental results demonstrate that the proposed MIL-GBLTM model outperforms other existing models by achieving the highest accuracy of 96.55% with a k-fold of 10, utilizing the wacv2016 dataset. Full article

Time-Sensitive Networking (TSN) has emerged as a critical communication protocol for automotive Ethernet to support the high-bandwidth, real-time, and deterministic transmission requirements of next-generation in-vehicle networks. However, a clear and effective TSN mechanism combination tailored to the mixed and bursty traffic characteristics of automotive scenarios remains lacking. To address this issue, this paper proposes a combined TSN scheduling mechanism for automotive scenarios. The highest-priority traffic is scheduled by class-based Time-Aware Shaper (TAS), periodic bursty sensor traffic is shaped by Credit-Based Shaper (CBS), and medium-priority traffic adopts Multi-Cyclic Queueing and Forwarding (MCQF). Based on Compositional Performance Analysis (CPA), this paper derives the worst-case latency upper bound expressions for CQF streams and optimizes the schedulability analysis to reduce conservative errors. Simulation verifies that the theoretically calculated bounds cover the maximum simulation latency, and the optimized analysis reduces conservatism, with peak conservative error of 3.07% in the ring scenario and 10.59% in the automotive scenario. Compared with the strict priority and TAMCQF (a combination of TAS and Multi-CQF), the proposed mechanism combination suppresses the latency jitter of mixed traffic, mitigates long-duration blocking of medium-priority traffic caused by high-priority burst data, and provides reliable deterministic transmission guarantees for automotive in-vehicle networks. Full article

Accurate waveform classification in noisy environments is an important task in modern communications, radar signal analysis, biomedical signal interpretation, industrial monitoring and other signal processing systems. This paper investigates the performance of three neural network architectures: Multilayer Perceptron (MLP), one-dimensional Convolutional Neural Network (1D-CNN), and Long Short-Term Memory (LSTM) for multiclass waveform classification in the presence of Additive White Gaussian Noise (AWGN). A time series dataset consisting of multiple waveform classes is generated and corrupted with AWGN across a wide range of signal-to-noise ratio (SNR) levels to simulate noisy signal distortion conditions. The three models are trained and evaluated under identical conditions to ensure a fair comparison. Their classification performance is evaluated in terms of accuracy, Confusion Matrix (CM), Receiver Operating Characteristic (ROC) curve and the Area Under the ROC curve (AUC) across varying SNR values. Simulation results demonstrate that the 1D-CNN effectively captures local temporal patterns and achieves superior robustness in classification at moderate and high SNR levels. The LSTM model demonstrates the ability to capture temporal dependencies in sequential waveform data but exhibits sensitivity to waveform variations due to amplitude, phase and frequency changes and noise at lower SNR values. The MLP, although computationally simpler, shows comparatively limited performance in low-SNR conditions due to its lack of temporal feature extraction capability. For the case of multiclass deterministic waveforms, the accuracy of classification for the 1D-CNN and LSTM is nearly 100% at SNR = 5 dB showing their robustness in classification, whereas the accuracy of MLP is approximately 70% that shows poor classification in noisy conditions. When there is random amplitude, frequency and phase variations in the waveforms, the accuracy of the 1D-CNN and MLP increases with SNR, and 1D-CNN superior to MLP. However, the LSTM accuracy fails to improve with SNR, resulting in poor classification performance in such a scenario. The results provide an insight into the suitability of different neural architectures for waveform classification tasks in noisy communication or other time series applications and highlight the advantages of convolutional feature extraction for robust signal recognition. Full article