Search Results (1,161)

This study provides a detailed comparative analysis of a three-hybrid intrusion detection method aimed at strengthening network security through precise and adaptive threat identification. The proposed framework integrates an Autoencoder-Gaussian Mixture Model (AE-GMM) with two supervised learning techniques, XGBoost and Logistic Regression, combining deep feature extraction with interpretability and stable generalization. Although the downstream classifiers are trained in a supervised manner, the hybrid intrusion detection nature of the framework is preserved through unsupervised representation learning and probabilistic modeling in the AE-GMM stage. Two benchmark datasets were used for evaluation: NSL-KDD, representing traditional network behavior, and UNSW-NB15, reflecting modern and diverse traffic patterns. A consistent preprocessing pipeline was applied, including normalization, feature selection, and dimensionality reduction, to ensure fair comparison and efficient training. The experimental findings show that hybridizing deep learning with gradient-boosted and linear classifiers markedly enhances detection performance and resilience. The AE–GMM-XGBoost model achieved superior outcomes, reaching an F1-score above 0.94 ± 0.0021 and an AUC greater than 0.97 on both datasets, demonstrating high accuracy in distinguishing legitimate and malicious traffic. AE-GMM-Logistic Regression also achieved strong and balanced performance, recording an F1-score exceeding 0.91 ± 0.0020 with stable generalization across test conditions. Conversely, the standalone AE-GMM effectively captured deep latent patterns but exhibited lower recall, indicating limited sensitivity to subtle or emerging attacks. These results collectively confirm that integrating autoencoder-based representation learning with advanced supervised models significantly improves intrusion detection in complex network settings. The proposed framework therefore provides a solid and extensible basis for future research in explainable and federated intrusion detection, supporting the development of adaptive and proactive cybersecurity defenses. Full article

The aim of this article is to develop a machine learning (ML)-based predictive model for industrial accidents in the energy sector. The dataset used in this study was obtained from the Kaggle platform and consists of summaries derived from reports of occupational incidents resulting in injuries or deaths between 2015 and 2017. A total of 4739 accident cases were included, containing information on accident date, accident summary, degree and nature of injury, affected body part, event type, human factors, and environmental factors. Six supervised machine learning models—Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, Gradient Boosting Decision Trees (GBDT), and Extreme Gradient Boosting (XGBoost)—were developed and compared to identify the most suitable model for the data. Model performance was evaluated using accuracy, precision, recall, F1-score, and the area under the receiver operating characteristic curve (AUC), which were selected to ensure reliable prediction in safety-critical accident scenarios. The results indicate that XGBoost and GBDT achieve superior performance in predicting human and environmental risk factors. These findings demonstrate the potential of machine learning for improving safety management in the energy sector by identifying risk mechanisms, enhancing safety awareness, and providing quantitative predictions of fatal and non-fatal accident occurrences for integration into safety management systems. Full article

Background: Cold chain distribution in Agri-Biotech supply chains faces serious challenges due to strict time windows, high temperature sensitivity, and conflict between different operational objectives, and conventional static approaches are unable to address these complexities. Methods: In this study, an integrated decision support framework is presented that combines multi-objective fuzzy modeling and an adaptive digital twin to simultaneously manage logistics costs, product quality degradation, and service time compliance under operational uncertainty. Key uncertain parameters are modeled using triangular fuzzy numbers, and the digital twin dynamically updates the decision parameters based on operational information. The proposed framework is evaluated using real industrial data and comprehensive computational experiments. Results: The results show that the proposed approach is able to produce stable and balanced solutions, provides near-optimal performance in benchmark cases, and is highly robust to demand fluctuations and temperature deviations. Digital twin activation significantly improves the convergence behavior and stability of the solutions. Conclusions: The proposed framework provides a reliable and practical tool for adaptive planning of cold chain distribution in Agri-Biotech industries and effectively reduces the gap between advanced optimization models and real-world operational requirements. Full article

Background: Port digital transition is central to competitiveness and sustainability, yet existing frameworks devoted to such transition toward smart port are descriptive, technology-centered, or weak on data governance. This study designs and empirically refines a comprehensive and novel ten-step roadmap relative to existing Port/Industry 4.0 models, synthesized from 14 partial frameworks that each cover only subsets of the transition, by considering data governance and consolidating cost, time, and impact in the selection step. Methods: We synthesized recent Industry 4.0 and smart port-related frameworks into a normalized sequence of steps embedded in the so-called roadmap, then examined it in an exploratory case of a technology deployment project in a Canadian port using stakeholder interviews and project documents. Evidence was coded with a step-aligned scheme, and stakeholder feedback and implementation observations assessed whether each step’s outcomes were met. Results: The sequence proved useful yet revealed four recurrent hurdles: limited maturity assessment, uneven stakeholder engagement, ad hoc technology selection and integration, and under-specified data governance. The refined roadmap adds a diagnostic maturity step with target-state setting and gap analysis, a criteria-based selection worksheet, staged deployment with checkpoints, and compact indicators of transformation performance, such as reduced logistics delays, improved energy efficiency, and technology adoption. Conclusions: The work couples theory-grounded synthesis with empirical validation and provides decision support to both ports and public authorities to prioritize investments, align stakeholders, propose successful policies and digitalization supporting programs, and monitor outcomes, while specifying reusable steps and indicators for multi-port testing and standardized metrics. Full article

Background and Objectives: Contrast-associated acute kidney injury (CA-AKI) is a frequent and clinically significant complication in patients with acute myocardial infarction (AMI) undergoing coronary angiography. Early and accurate risk stratification remains challenging with conventional models that rely on linear assumptions and limited variable integration. This study aimed to evaluate and compare the predictive performance of multiple machine learning (ML) algorithms with traditional logistic regression and the Mehran risk score for CA-AKI prediction and to explore key determinants of risk using explainable artificial intelligence methods. Materials and Methods: This retrospective, single-center study included 1741 patients with AMI who underwent coronary angiography. CA-AKI was defined according to KDIGO criteria. Multiple ML models, including gradient boosting machine (GBM), random forest (RF), XGBoost, support vector machine, elastic net, and standard logistic regression were developed using routinely available clinical and laboratory variables. A weighted ensemble model combining the best-performing algorithms was constructed. Model discrimination was assessed using area under the receiver operating characteristic curve (AUC), along with sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Model interpretability was evaluated using feature importance and SHapley Additive exPlanations (SHAP). Results: CA-AKI occurred in 356 patients (20.4%). In multivariable logistic regression, lower left ventricular ejection fraction, higher contrast volume, lower sodium, lower hemoglobin, and higher neutrophil-to-lymphocyte ratio (NLR) were independently associated with CA-AKI. Among ML approaches, the weighted ensemble model demonstrated the highest discriminative performance (AUC 0.721), outperforming logistic regression and the Mehran risk score (AUC 0.608). Importantly, the ensemble model achieved a consistently high NPV (0.942), enabling reliable identification of low-risk patients. Explainability analyses revealed that inflammatory markers, particularly NLR, along with sodium, uric acid, baseline renal indices, and contrast burden, were the most influential predictors across models. Conclusions: In patients with AMI undergoing coronary angiography, interpretable ML models, especially ensemble and gradient boosting-based approaches, provide superior risk stratification for CA-AKI compared with conventional methods. The high negative predictive value highlights their clinical utility in safely identifying low-risk patients and supporting individualized, risk-adapted preventive strategies. Full article

Accurate and interpretable detection of pulmonary emphysema on low-dose computed tomography (LDCT) remains a critical challenge for large-scale screening and population health studies. This work proposes a quality-controlled and interpretable deep learning pipeline for emphysema assessment using ResNet-152 embeddings. The pipeline integrates automated lung segmentation, quality-control filtering, and extraction of 2048-dimensional embeddings from mid-lung patches, followed by analysis using logistic regression, LASSO, and recursive feature elimination (RFE). The embeddings are further fused with quantitative CT (QCT) markers, including %LAA, Perc15, and total lung volume (TLV), to enhance robustness and interpretability. Bootstrapped validation demonstrates strong diagnostic performance (ROC-AUC = 0.996, PR-AUC = 0.962, balanced accuracy = 0.931) with low computational cost. The proposed approach shows that ResNet embeddings pretrained on CT data can be effectively reused without retraining for emphysema characterization, providing a reproducible and explainable framework suitable as a research and screening-support framework for population-level LDCT analysis. Full article

This paper addresses the optimization challenges in urban logistics with the aim of enhancing the sustainability of last-mile delivery. By focusing on the collaborative delivery between unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs), we propose a novel approach to reducing energy consumption and operational inefficiencies. A bilevel mixed-integer linear programming (Bilevel-MILP) model is developed, integrating road network topology with dynamic energy constraints. Departing from traditional single-delivery modes, the paper establishes a multi-task continuous delivery framework. By incorporating a dynamic charging point selection strategy and path–energy coupling constraints, the model effectively mitigates energy limitations and the issue of repeated returns for UAV charging in complex urban road networks, thereby promoting more efficient resource utilization. At the algorithmic level, a Collaborative Delivery Path Optimization (CDPO) framework is proposed, which embeds an Improved Sparrow Search Algorithm (ISSA) with directional initialization and a Hybrid Genetic Algorithm (HGA) with specialized crossover strategies. This enables the synergistic optimization of UAV delivery sequences and UGV charging decisions. The simulation results demonstrate that, in scenarios with a task density of 20 per 100 km², the proposed CDPO algorithm reduces the total delivery time by 33.9% and shortens the UAV flight distance by 24.3%, compared to conventional fixed charging strategies (FCSs). These improvements directly contribute to lowering energy consumption and potential emissions. The road network discretization approach and dynamic candidate charging point generation confirm the method’s adaptability in high-density urban environments, offering a spatiotemporal collaborative optimization paradigm that supports the development of sustainable and intelligent urban logistics systems. The obtained results provide practical insights for the design and deployment of efficient UAV–UGV collaborative logistics systems in urban environments, particularly under high-task-density and energy-constrained conditions. Full article

The integration of biomethane into the natural gas infrastructure is a critical element of energy-sector decarbonization, yet optimal infrastructure development scenarios remain insufficiently compared using unified decision frameworks. This study evaluates three biomethane market integration scenarios—direct connection to the gas system, biomethane injection points (compressed biomethane transported by trucks to the gas system), and off-grid delivery using the multi-criteria decision-making method TOPSIS. Environmental, economic, and technical dimensions are jointly assessed. Results indicate that direct connection to the system provides the most balanced overall performance, achieving the highest integrated score (Ci = 0.70), driven by superior environmental and technical characteristics. Biomethane injection points demonstrate strong economic advantages (Ci = 0.49), particularly where capital investments need to be reduced or there is limited access to the gas system, but show weaker environmental and technical performance. Off-grid solutions perform poorly in integrated assessment (Ci = 0.00), reflecting limited scalability and high logistical complexity, although niche applications may remain viable under specific conditions. Sensitivity analysis confirms the robustness of these rankings across a wide range of weighting assumptions, strengthening the reliability of the findings for policy and infrastructure planning. This study provides one of the first integrated multi-criteria assessments explicitly incorporating virtual pipeline logistics, offering a transferable decision-support framework for sustainable biomethane development in diverse regional contexts. Full article

Background. Rheumatoid arthritis (RA) is a chronic autoimmune disease frequently accompanied by cardiovascular, respiratory, skeletal, psychiatric, and neoplastic comorbidities that are associated with higher morbidity and poorer health-related quality of life (HRQoL). This study evaluated the associations between comorbidities and patient-reported physical health, emotional distress, daily functioning, and social relationships in adults with RA and explored patient-reported unmet needs relevant to integrated care. Methods. We conducted a cross-sectional survey among 286 adults with physician-confirmed RA, using a structured questionnaire (ICRA-Q) administered between June and July 2025 via online platforms and in-hospital supervised completion. The survey captured demographics, patient-reported physician-diagnosed comorbidities (current and/or past), perceived disease impact, functional limitations, emotional and social consequences, access to treatment, financial burden, and support needs. Analyses included descriptive statistics, χ² tests, t-tests/ANOVA, effect sizes (Cramer’s V and standardized mean differences), and multivariable logistic regression to explore predictors of high HRQoL impact and high difficulty in disease management. An exploratory classification into high-risk phenotypes was performed using predefined clinical, psychological, and socioeconomic criteria. Results. Most participants (98.6%) reported at least one comorbidity, most commonly hypertension, osteoporosis, and cardiovascular disease. Higher comorbidity burden and depression/anxiety were strongly associated with higher pain, reduced mobility, emotional distress, and financial strain. Exploratory high-risk phenotypes (severe somatic multimorbidity, high psychological vulnerability, high socioeconomic burden, and a composite very high-risk profile) were associated with poorer HRQoL indicators. Younger age, shorter disease duration, and higher perceived social support were associated with lower perceived burden. Conclusions. In this cross-sectional, patient-reported study, comorbidity burden—particularly psychological comorbidity—was strongly associated with poorer HRQoL and greater management difficulty in RA. These findings support the need for multidisciplinary, integrated care pathways; however, subgroup phenotypes should be considered exploratory and require external validation. Full article

Background: The persistently low 5-year survival rate for pancreatic cancer (PC) underscores the critical need for early detection. However, population-wide screening remains impractical. Artificial Intelligence (AI) models using electronic health record (EHR) data offer a promising avenue for pre-symptomatic risk stratification. Objective: To systematically review and meta-analyze the performance of AI models for PC prediction based exclusively on structured EHR data. Methods: We systematically searched PubMed, MedRxiv, BioRxiv, and Google Scholar (2010–2025). Inclusion criteria encompassed studies using EHR-derived data (excluding imaging/genomics), applying AI for PC prediction, reporting AUC, and including a non-cancer cohort. Two reviewers independently extracted data. Random-effects meta-analysis was performed for AUC, sensitivity (Se), and specificity (Sp) using R software version 4.5.1. Heterogeneity was assessed using I² statistics and publication bias was evaluated. Results: Of 946 screened records, 19 studies met the inclusion criteria. The pooled AUC across all models was 0.785 (95% CI: 0.759–0.810), indicating good overall discriminatory ability. Neural Network (NN) models demonstrated a statistically significantly higher pooled AUC (0.826) compared to Logistic Regression (LogReg, 0.799), Random Forests (RF, 0.762), and XGBoost (XGB, 0.779) (all p < 0.001). In analyses with sufficient data, models like Light Gradient Boosting (LGB) showed superior Se and Sp (99% and 98.7%, respectively) compared to NNs and LogReg, though based on limited studies. Meta-analysis of Se and Sp revealed extreme heterogeneity (I² ≥ 99.9%), and the positive predictive values (PPVs) reported across studies were consistently low (often < 1%), reflecting the challenge of screening a low-prevalence disease. Conclusions: AI models using EHR data show significant promise for early PC detection, with NNs achieving the highest pooled AUC. However, high heterogeneity and typically low PPV highlight the need for standardized methodologies and a targeted risk-stratification approach rather than general population screening. Future prospective validation and integration into clinical decision-support systems are essential. Full article

The article presents a method that supports the evaluation of design manufacturability in the area of input material selection, enabling the reduction in material diversity under single-piece and small-batch production conditions. The proposed approach combines the analysis of alternative materials with activity-based costing (ABC) and data concerning actual and planned material requirements. The method enables the assessment of the impact of semi-finished product substitution on material costs, processing costs, production organisation, and material-management costs before order execution is launched. In the conducted case study, it was demonstrated that effective management of material diversity can significantly reduce the range of materials and decrease total manufacturing costs. For the analysed period, the number of material items was reduced from 32 to 19 (a 41% reduction), resulting in cost savings of approximately 11,000 PLN. In addition to total cost, the approach supports the assessment of operational benefits associated with reduced material diversity, such as a lower number of material items, fewer suppliers, reduced inbound inspection and receipt operations, and decreased inventory levels and capital tied up in stock. Material substitution may decrease or increase direct material costs and may increase machining time when larger dimensions are used; therefore, the method jointly evaluates cost and lead-time impacts prior to order release. The results confirm that integrating design, technological, and logistics data is an effective approach to rationalising material management in machinery manufacturing enterprises. Full article

With rapid urbanization and population growth, efficient transportation systems are increasingly crucial, particularly in sectors like healthcare and pharmaceutical logistics, which face unique challenges. In Morocco, there is a lack of studies on pharmaceutical transport, especially regarding costs and delivery conditions, creating a need for a specialized model. This research presents the development and validation of a predictive model for optimizing urban transport in Morocco. Tested across key sectors—pharmaceuticals, agri-food, electronics, and manufactured goods—the model demonstrated strong performance, though variations emerged based on product complexity. Notably, the agri-food sector presented greater logistical challenges, while the manufacturing and electronics sectors yielded higher prediction accuracy. By integrating statistical process control (SPC) and Lean Six Sigma principles, the model ensures ongoing performance monitoring and continuous improvement. It supports cost reduction, time optimization, and lower environmental impact through enhanced route planning and delivery efficiency. The pharmaceutical sector was selected as a case study due to its critical logistical constraints, such as cold chain requirements and the need for high reliability. Python was used for model development, enabling rapid iteration and collaborative validation. The results confirm the model’s adaptability and generalizability to similar urban environments across North and Sub-Saharan Africa. The study offers a robust and scalable framework for improving transport efficiency while aligning with sustainability and smart mobility goals. Full article

Background: Even though there have been improvements in antimicrobial and supportive therapies, sepsis and septic shock are still major causes of death in intensive care units. Early prognostic stratification is very important for helping doctors make decisions. Platelet-derived indices may provide useful, low-cost indicators that signify both inflammatory activation and coagulation irregularities. This study looked at how well different platelet-based ratios could predict death in the hospital from sepsis. Materials and Methods: We performed a prospective observational study spanning one year in a tertiary ICU, enrolling 114 adult patients diagnosed with sepsis or septic shock. Upon admission, four platelet-related biomarkers were measured: the C-reactive protein-to-platelet ratio (CPR), the platelet-to-lymphocyte ratio (PLR), the platelet-to-white blood cell ratio (PWR), and the platelet-to-creatinine ratio (PCR). Logistic regression models and receiver operating characteristic (ROC) analyses were employed to assess predictive accuracy. Results: Compared to survivors, non-survivors (n = 39) had much higher CRP levels and CPR values, alongside lower platelet and lymphocyte counts. The CPR index showed the best ability in differentiating between non-survivors and survivors (AUC 0.757), with a best cutoff of 0.886. In simplified multivariate models, CPR was still an independent predictor of death in the hospital (OR 1.98; 95% CI 1.22–3.21), whereas PLR and PWR were not. PCR showed a non-significant trend toward lower values in not survivors. Conclusions: CPR is a strong and clinically viable predictor of early mortality in sepsis, outperforming other platelet-based indices. Derived from routine laboratory parameters, CPR serves as a valuable adjunct for initial risk stratification in the ICU. To further confirm its prognostic role and incorporation into current scoring systems, large-scale multicenter studies with longitudinal measurements are warranted to validate its prognostic utility and integration into existing scoring systems. Full article

Background and Objectives: Coronary artery disease (CAD) is a leading cause of morbidity and mortality worldwide. An early and accurate diagnosis is essential for effective clinical management and risk stratification. Recent advances in machine learning (ML) have provided opportunities to enhance the diagnostic performance by integrating multidimensional patient data. This study aimed to develop and compare several supervised ML algorithms for early CAD diagnosis using demographic, anthropometric, biochemical, and psychosocial parameters. Materials and Methods: A total of 300 adult patients (165 CAD-positive and 135 controls) were retrospectively analyzed using a dataset comprising 21 biochemical markers, body composition metrics, and self-reported eating behavior scores. Six ML algorithms, k-nearest neighbors (k-NNs), support vector machines (SVMs), artificial neural networks (ANNs), logistic regression (LR), naïve Bayes (NB), and decision trees (DTs), were trained and evaluated using 10-fold cross-validation. Model performance was assessed based on accuracy, sensitivity, false-negative rate, and area under the Receiver Operating Characteristic (ROC) curve (AUC). Results: The k-NN model achieved the highest performance, with 98.33% accuracy and an AUC of 0.99, followed by SVM (96.67%, AUC = 0.95) and ANN (95.33%, AUC = 0.98). Patients with CAD exhibited significantly higher levels of glucose, triglycerides (TGs), LDL cholesterol (LDL-C), and abdominal obesity, while vitamin B12 levels were lower (p < 0.001). Although emotional and mindful eating scores differed significantly between the groups, their contribution to model performance was limited. Conclusions: Machine learning models, particularly k-NN, SVM, and ANN, have demonstrated high accuracy in distinguishing CAD patients from healthy controls when applied to a diverse set of clinical and behavioral variables. This study highlights the potential of integrating psychosocial and clinical data to enhance CAD prediction models beyond traditional biomarkers. Full article

Background/Objectives: Gestational exposure to substances (GES) is associated with adverse developmental outcomes. Early identification is limited by reliance on self-reported data. This study assessed the incidence and predictors of discordance in GES reporting between birth certificates and Medicaid claims among Medicaid-covered births in Nevada from 2022 to 2024. Methods: A statewide, hospital-clustered, cross-sectional analysis was conducted using linked Medicaid billing and birth record data. Discordance was defined as GES identified in one source but not the other. Incidence per 1000 live births was stratified by demographic characteristics. Multilevel logistic regression assessed patient- and hospital-level predictors, with random hospital intercepts. Results: Among 50,394 live births, the discordance rate was 95.09 per 1000 (95% Confidence Interval: 92.5–97.7). Substantial disparities were observed by race/ethnicity, socioeconomic status, and geography, with higher discordance among White non-Hispanic mothers, those residing in rural or frontier counties, and individuals with lower educational attainment or living in lower-income areas. Modest but meaningful variation was also observed across hospitals, including differences by hospital size and teaching or research status. Conclusions: Findings highlight substantial discordance in GES reporting and underscore the limitations of single-source surveillance. Findings also have clear policy relevance, indicating that improved cross-system data integration would strengthen statewide surveillance, enhance early detection, and support more equitable resource allocation and intervention strategies. Full article