Search Results (40,033)

The growing energy demand in rural areas such as the ejido Boca del Cerro, located in Tenosique, Tabasco (Mexico), near the Usumacinta River, calls for sustainable energy solutions such as microgrids. This study proposes an energy management system combining renewable energy forecasting and fuzzy control for a simulated small autonomous rural microgrid scenario designed to supply a fixed priority load of 5 kW and a variable flexible load ranging from 1 to 10 kW. Three LSTM architectures (vanilla, stacked, and bidirectional) are compared for predicting solar irradiance, wind speed, and river flow. The vanilla model is optimized using Hyperband to improve prediction accuracy, particularly for flow rate, which is rarely addressed in similar studies. Forecasts feed into models of photovoltaic, wind, and hydro systems within the microgrid. Energy dispatch is managed through fuzzy logic control. The fuzzy controller supports load prioritization, battery charge/discharge management, and surplus energy redirection to an absorbing load. The final vanilla LSTM achieved RMSE values of 25.741, 0.302, and 12.644 for solar irradiance, wind speed, and river flow, respectively, with NSE values above 0.949 in all cases. These results indicate high forecasting accuracy for solar irradiance and river flow, with limited improvement for wind speed. Overall, the proposed EMS enables effective energy flow management, while the integration of hydrokinetic turbines with AI-based forecasting represents a novel contribution. Full article

The discharge of metal-complex dyes from textile industries poses significant environmental challenges due to their chemical stability and resistance to conventional biological treatment. This study examined the degradation of Acid Black 194 (AB–194), a 1:2 chromium-complex azo dye, using Co²⁺-activated peroxymonosulfate (PMS). A central composite design based on response surface methodology was used to evaluate the effects of Co²⁺ (5.93–20.07 µM), PMS (1.67–7.33 mM), and dye (13.79–56.21 mg L⁻¹) concentrations on decolorization and mineralization. The polynomial models demonstrated strong predictive accuracy (R² > 0.9896), identifying Co²⁺ and dye concentrations as the most influential factors. Under optimal conditions (18.0 µM Co²⁺, 6.5 mM PMS, 20.0 mg L⁻¹ dye), 99.19% decolorization was achieved at 30 min and 41.43% TOC removal at 240 min. Degradation kinetics were described by a mechanistic model incorporating 15 elementary reactions that comprise the Co²⁺/Co³⁺ redox cycle, radical generation, and dye oxidation, yielding a global R² of 0.9617. Estimated rate constants for dye oxidation (k₁₄ = 3.52 × 10⁹ M^–1 s^–1 for and k₁₅ = 2.00 × 10¹⁰ M^–1 s^–1 ) were consistent with values reported for aromatic compounds in sulfate radical systems. Radical contribution analysis confirmed sulfate radicals as the principal oxidizing species, accounting for 96.75% of the overall process. Full article

Introduction: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is widely used for diagnosing pulmonary diseases causing mediastinal lymphadenopathy. However, non-diagnostic results may occur. This study investigated factors associated with non-diagnostic cytological results in EBUS-TBNA. Methods: This retrospective study included patients who underwent EBUS-TBNA at a tertiary hospital between March 2019 and December 2023. Data on demographics, biopsy techniques, cyto-/histopathological results, sonographic lymph node measurements, and pre-procedural PET-CT SUVmax values were recorded. Cytological results were classified as diagnostic or non-diagnostic. We analyzed the characteristics and associated factors of patients who were non-diagnostically identified. Results: Among 776 patients undergoing EBUS-TBNA, 502 (64.7%) were male, with a mean age of 61.5 ± 12.6 years. A total of 1110 lymph nodes were sampled. Of the patients, 14.1% had a non-diagnostic cytology. Among the diagnosed patients, cytological findings showed 58.9% non-malignant, 41.1% malignant. The most sampled station was station 7 (72.9%), with an average of 5.9 ± 1.4 aspirations. Diagnostic cases had significantly more aspirations (p = 0.022) and sampled larger lymph node sizes (p < 0.001). Each 1 mm increase in lymph node size raised the likelihood of diagnostic results by 1.04 times (adjOR = 1.04, 95% CI = 1.02–1.08, p = 0.002). The largest lymph node size significantly predicted diagnostic results (AUROC = 0.611, p < 0.001). A cut-off of 19.55 mm had 67.0% sensitivity and 52.2% specificity. Conclusion: Sampled larger lymph nodes increase diagnostic yield in EBUS-TBNA, reducing the need for repeat procedures and enabling earlier treatment, thereby decreasing morbidity and mortality. Full article

Objectives: To validate claims-based algorithms for identifying thyroid eye disease (TED) cases and assess whether machine learning can improve case identification in a large, linked real-world dataset. Methods: Using a large, linked database from Komodo Health^® and Academy IRIS^® Registry, we evaluated six rule-based algorithms incorporating Graves’ disease (GD), eye symptoms and signs. The IRIS Registry’s curated data, based on confirmed TED diagnoses from medical notes, served as the reference standard. Additionally, we developed supervised machine learning models using demographic, diagnostic, procedural, and medication data. Feature selection was performed using recursive feature elimination to rank predictive codes and construct a simplified, interpretable model. Cross-validation was used to assess model performance and compare performance with the rule-based algorithms. Results: The rule-based algorithms demonstrated a trade-off between sensitivity and specificity, with some achieving high specificity but limited sensitivity. Algorithm 1 had the highest sensitivity (48.7%) but lower specificity (59.9%) and PPV (75.8%). Algorithms 2–5 demonstrated higher specificity (87.2–93.5%) but lower sensitivity (17.8–27.0%). Algorithm 6 improved sensitivity (33.4%) compared to Algorithms 2–5 while maintaining high specificity (86.8%) and a strong PPV (86.7%). Machine learning models demonstrated similar trade-offs. One model achieved improved specificity (77.2%) with sensitivity of 49.3%, outperforming Algorithm 1 in specificity while matching its sensitivity. Another model maximized specificity (91.7%) and PPV (89.8%) at a reduced sensitivity of 28.5%. These results highlight the flexibility of machine learning models in adjusting performance to address different research objectives. Conclusions: This study evaluated existing rule-based algorithms for identifying TED cases in claims data, revealing trade-offs between sensitivity and specificity. Machine learning models provide additional flexibility, allowing performance to be tailored to specific research use cases. While no single method consistently outperformed others across all metrics, both rule-based and machine learning approaches demonstrated value in improving TED case identification using real-world data sources. Full article

To improve the efficiency of comparative process-window screening in selective laser melting (SLM), this study developed a finite element simulation-driven machine learning framework for 316L stainless steel. A simulation dataset covering laser power (LP), scanning speed (SS), heat-source diameter (HSD), and substrate preheating temperature (SPH) was generated using ANSYS and used to train nine regression models. In the present work, the primary machine learning target was defined as the simulated average thermal stress, σ_avg, which is used as a simulation-derived comparative thermal stress indicator for ranking process conditions within the investigated parameter window rather than as a direct prediction of the final residual-stress field. Among the evaluated models, the Backpropagation Neural Network (BPNN) showed the best predictive performance and was selected as the representative surrogate model because of its strong predictive accuracy, stable behavior, and direct applicability to the present structured tabular dataset. Shapley additive explanations (SHAP) and partial dependence plots (PDPs) indicated that LP is the dominant variable governing the σ_avg-based response, followed by SPH, whereas SS and HSD mainly affect the response through secondary or coupled effects. Within the investigated parameter window, conditions near 180–200 W corresponded to a relatively lower predicted σ_avg level. Experimental observations provided limited but meaningful trend-level support for the simulation-guided screening results: metallographic examination showed improved forming quality near 200 W, while XRD-derived macroscopic stress estimates exhibited a similar variation trend to the simulated σ_avg values under the tested LP–SS conditions. These results suggest that the proposed framework can serve as an efficient surrogate-based tool for comparative parameter screening in SLM-fabricated 316L stainless steel within the assumptions and parameter range of the present model. Full article

Background/Objectives: Pneumonia remains a leading cause of emergency department visits worldwide, requiring rapid and objective risk stratification. While traditional scoring systems like CURB-65 and the Pneumonia Severity Index (PSI) are well-established, there is a constant need for dynamic biomarkers reflecting the underlying pathophysiology. This study aims to investigate the prognostic value of the hemoglobin-to-lactate ratio (HLR) in predicting mortality among pneumonia patients. Methods: This retrospective cohort study included 183 adult patients diagnosed with pneumonia at a tertiary training and research hospital between October 2024 and November 2025. Demographic data, clinical findings, laboratory parameters, and prognostic scores (CURB-65, PSI) were recorded. The impact of HLR on mortality was evaluated using univariate and multivariate logistic regression, while its predictive performance was assessed via Receiver Operating Characteristic (ROC) analysis and compared with clinical scores using DeLong’s method. Results: The overall mortality rate was 32.8%. HLR values were significantly lower in the exitus group compared to survivors (4.68 vs. 6.92, p < 0.001). Multivariate analysis revealed that an HLR ≤ 5.65 was an independent predictor of mortality, associated with a 10-fold increase in risk (OR: 10.0; 95% CI: 4.15–24.19; p < 0.001). HLR demonstrated high predictive power (AUC = 0.802), comparable to CURB-65 (AUC = 0.807) and PSI (AUC = 0.829). Notably, the combined HLR + CURB-65 model provided the highest diagnostic accuracy (AUC = 0.857, p = 0.037). Conclusions: HLR is a low-cost and easily accessible biomarker for predicting mortality in pneumonia. It effectively reflects the physiological balance between tissue oxygenation and metabolic failure. Integrating HLR into clinical practice, particularly when combined with traditional scores, can enhance risk (decision of discharge, admission unit [ward, ICU], evaluation of prognosis) in the emergency department. Full article

Cervical cancer screening with high-risk human papillomavirus (HR-HPV) testing requires effective triage of HPV-positive women. Koilocytosis is a classic cytopathic effect of HPV infection, but its clinical significance in low-grade squamous intraepithelial lesions (LSILs) remains unclear. We retrospectively evaluated 157 HPV-positive women with LSIL cytology and follow-up data, including 140 women with concurrent biopsy results. Koilocytes were identified in 93/157 cases (59.2%) and were less frequent in HPV16/18-positive cases. Cervical intraepithelial neoplasia ≥ grade 2 (CIN2+) was detected in 9/84 koilocyte-positive cases (10.7%) and 16/56 koilocyte-negative cases (28.6%), whereas non-CIN findings were more common in koilocyte-positive cases. Koilocyte-positive cases also showed a longer time to regression from LSIL to negative for intraepithelial lesions or malignancy. These findings suggest that koilocytosis mainly reflects productive HPV infection and has limited utility for predicting CIN2+ in HPV-based screening triage. Excluding koilocytosis-driven low-grade cytological changes from triage positivity criteria may improve specificity and positive predictive value, supporting higher triage thresholds. Full article

With the large-scale grid integration of renewable energy sources such as wind power and photovoltaics, power system net load fluctuations have become significantly more severe, imposing higher demands on system flexibility. Traditional optimal transmission switching (OTS) models require the simultaneous optimization of continuous and discrete variables, resulting in high computational complexity that renders them unsuitable for daily real-time scheduling in large-scale power systems. This paper develops a flexible real-time rolling optimization scheduling model that incorporates OTS and proposes a two-stage fast solution framework based on deep neural networks (DNN). In the offline training phase, a multilayer perceptron-based DNN is trained using load and renewable generation data to rapidly and accurately predict the optimal line switching scheme. In the online application phase, the network topology predicted by the DNN transforms the original mixed-integer linear programming problem into a standard linear programming problem, substantially reducing computational complexity and solution time. Case studies on the modified IEEE 118-bus and IEEE 300-bus systems show that the proposed method achieves high prediction accuracy, reduces solution time by up to 117 times, and maintains nearly identical system operating costs to the physics-driven approach in the majority of cases. The results demonstrate that the proposed approach effectively balances computational efficiency and economic performance, verifying the practical value of optimal transmission switching in enhancing large-scale renewable energy accommodation and overall power system flexibility. Full article

Competition for water resources makes it necessary to advance research focused on estimating the water needs of row crops, such as vineyards. Following the FAO-56 methodology and the A&P approach, the soil water balance model was applied to a vineyard with continuous vegetation cover in temperate climate conditions (Galicia, Spain). Basal crop coefficients adjusted to local conditions were obtained for both the vineyard and the active vegetation. After SIMDualKc model adjustment, r² values greater than 0.86 were obtained, along with goodness-of-fit indicators that demonstrate the model’s ability to predict soil water content (PBIAS_avg = 1.16; EF_avg = 0.89; dIA_avg = 0.97). A correction factor is proposed that improves the partitioning of the transpiration component in row crops with active cover. The transpiration demand of the vineyard increased by 35% in four study cases (northern Portugal, northwestern Spain, and Italy). The proposed correction factor is shown to be in line with the actual conditions and complex behaviour of a vineyard with active vegetation cover, which opens the way for improved water requirement prediction in complex management situations such as the one studied here. The proposed methodology is expected to improve the efficiency of irrigation management through more accurate determination of the real water amount required by orchards. Full article

This study develops an efficient fatigue prediction framework for offshore wind turbine (OWT) monopile foundations under coupled wind–wave conditions using four surrogate models: XGBoost, Random Forest (RF), Support Vector Regression (SVR), and Gaussian Process Regression (GPR). A finite element model (FEM) incorporating soil–pile interaction is established to accurately capture structural responses under realistic environmental loading. Fatigue damage is evaluated through time-domain simulations based on this model. A surrogate modeling approach is employed to capture the nonlinear mapping between environmental variables and fatigue damage using 60 representative samples. Results show that the proposed framework significantly improves computational efficiency while maintaining predictive reliability. Among the models evaluated, GPR yields the highest prediction accuracy, while SVR shows comparable performance. In contrast, XGBoost and RF exhibit relatively larger deviations. Parametric analysis reveals that fatigue damage is positively correlated with wind speed and significant wave height, but inversely correlated with peak wave period. Further, wind-induced loading dominates fatigue accumulation, and conventional load superposition methods underestimate fatigue damage due to nonlinear wind–wave coupling effects. Furthermore, fatigue damage exhibits pronounced circumferential variation, with maximum values occurring in the fore-aft directions. Full article

This study proposes an explainable framework for ESG portfolio rebalancing by integrating carbon credit signals, technical indicators, and Transformer-inspired forecasting into a unified decision process. The investable universe consists of six ESG-themed ETFs, namely ESGU, SUSA, ICLN, TAN, KRBN, and KGRN. Carbon-related sustainability information is represented by four S&P carbon indices, including GCC, CCA, EUA, and UCITS. Within the proposed framework, Transformer, Informer, and Temporal Fusion Transformer are used to predict next-day returns, and the forecast outputs are translated into portfolio decisions through threshold filtering, Softmax-based allocation, and inertia smoothing under fixed transaction costs. The empirical results show that the proposed framework remains competitive against Equal Weight, Risk Parity, and Momentum benchmarks, although its advantage is conditional rather than uniformly dominant across all metrics. Informer delivers the strongest Sharpe ratio among the model-based strategies, while Transformer exhibits a more stable risk profile. The ablation results indicate that the smoothing mechanism has the clearest effect on turnover and allocation stability, whereas the incremental value of carbon-related inputs is most visible in Informer. The uncertainty assessment further shows that many benchmark differences are not consistently significant under repeated resampling, but the performance weakening caused by removing carbon inputs in Informer remains identifiable. The subperiod analysis shows that benchmark rules are more competitive in 2024H1, whereas model-based strategies gain relative strength in 2024H2. The explainability analysis indicates that carbon-feature contributions are concentrated more strongly in Intermediate and Carbon-Sensitive asset groups and remain weaker in Broad ESG assets; feature-level and SHAP beeswarm evidence further shows that the three architectures rely on GCC, CCA, EUA, and UCITS in different ways. These findings suggest that carbon-related sustainability signals can provide economically meaningful allocation information in selected settings when they are combined with suitable model architecture and disciplined rebalancing control, thereby supporting a competitive and explainable ESG portfolio rebalancing framework. Full article

Background/Objectives: Community-acquired pneumonia (CAP) is a leading cause of mortality in older adults. Traditional prognostic scores may underestimate risk in frail patients by assuming linear relationships between predictors and outcomes. This study aimed to develop and validate explainable machine learning models integrating the administrative Hospital Frailty Risk Score (HFRS) to predict in-hospital mortality in a nationwide cohort of older adults in Chile. Methods: A retrospective cohort study was conducted using anonymized hospital discharge records from the Chilean National Health Fund (FONASA), including 58,306 hospitalization episodes of adults aged ≥60 years across 72 public hospitals. Fourteen supervised machine learning algorithms were trained using five routinely collected predictors: age, sex, HFRS, Charlson Comorbidity Index, and length of stay. Model performance was evaluated on an independent test set using AUC-ROC. SHAP (SHapley Additive exPlanations) values were calculated to assess global and individual predictor contributions. Results: The Extra Trees classifier achieved the highest discriminative performance (AUC-ROC 0.862), outperforming logistic regression (0.642) and other linear models. SHAP analyses identified HFRS as the most influential predictor (mean |SHAP| = 0.66), followed by length of stay, age, and comorbidities. Conclusions: Ensemble tree-based models incorporating administrative frailty measures provide superior mortality prediction compared to traditional linear approaches. Frailty emerged as the primary driver of risk, supporting scalable early stratification using routinely available hospital data. Full article

Background/Objectives: Carbonic anhydrase I (CAI) is a zinc-dependent metalloenzyme whose inhibitor discovery requires both effective navigation of chemical space and explicit evaluation of coordination-credible binding hypotheses. We aimed to develop an interpretable and reproducible QSAR-to-structure workflow for CAI inhibitor discovery. The workflow links potency prediction with zinc-site plausibility and early developability to support decision-oriented prioritization of new CAI inhibitor candidates. Methods: CAI inhibitors were retrieved from ChEMBL (CHEMBL261) and modeled as pK_i = 9 – log₁₀(K_i[nM]). AlvaDesc v3.0.8 generated 4224 2D descriptors, which were reduced using train-only preprocessing, variance filtering, correlation pruning, and bagged-tree ranking to a top-100 panel. Five regressors (elastic net, CART, bagging, GB, and XGB) were benchmarked on a held-out test set. Potent ChEMBL seeds (K_i ≤ 10 nM) were used for a 90% 2D similarity PubChem expansion. Predicted hits were then externally validated using independently available PubChem CAI K_i records. Ten novel candidates lacking CAI K_i data were docked to CAI (PDB: 1AZM) via SwissDock AutoDock Vina in neutral and relevant anionic states, with pose selection constrained by a Zn-donor filter (Zn-N/O ≤2.6 Å). SwissADME was used to profile physicochemical space, alerts, and absorption/distribution proxies. Results: The bagging model showed the best test generalization (R² = 0.646; RMSE = 0.61; MAE = 0.45). PFI and SHAP converged on sulfur/heteroatom connectivity and polar–lipophilic organization as dominant potency drivers. PubChem expansion yielded 25,315 analogs and 233 candidates at predicted pK_i ≥ 8.0; external validation on 145 CAI-measured hits gave R² = 0.358 (RMSE = 0.456; MAE = 0.320). Across 20 ligand/protomer docking runs, 12 produced canonical Zn-anchored poses (10 Zn-N; 2 Zn-O). SwissADME indicated consensus logP values from −0.65 to 3.21, 0/10 PAINS alerts, and predominantly favorable drug-likeness (8/10 with zero Lipinski violations), supporting tiered advancement. Conclusions: Integrating interpretable QSAR, external PubChem validation, coordination-aware docking, and SwissADME yields a practical triage framework for CAI inhibitor discovery. The resulting tiered shortlist identifies two Zn-N-anchored N-alkyl sulfamides (CIDs 103935964 and 112684680) and one Zn-O-anchored carboxylate control (CID 122367674) as highest-priority computational hypotheses for staged biochemical evaluation. Full article

Yellow onion (Allium cepa L.) outer skins are a high-volume agricultural waste that can be converted into commercially valuable bioproducts using various extraction techniques. This research focused on optimizing a green ultrasound-assisted extraction (UAE) method which allows for the isolation of several phytochemicals valued for their health benefits, such as polyphenols and flavonoids. HPLC/UV analysis of the extracts showed that the main component was quercetin. A one-factor-at-a-time (OFAT) design was used to identify the extraction parameters needed in order to maximize the amount of extracted target phytochemicals. The polyphenols, flavonoids and quercetin contents, along with the antioxidant activity of the extracts, were optimized by response surface methodology using a Box–Behnken design. Ultrasound amplitude, ethanol concentration, and time were selected as the most appropriate variables. The final results showed that TPC ranged from 78.16 to 97.16 mg GAE/g DM, TFC ranged from 22.77 to 26.46 mg QE/g DM, while CUPRAC values varied between 145.24 and 163.75 mg TE/g DM. The optimal extraction conditions were determined using a Box–Behnken model as 30% ultrasound amplitude, 53% ethanol concentration, and an extraction time of 13 min. The use of these conditions allowed the TPC, TFC and CUPRAC to show predicted values of 97.8 mg GAE/g DM, 27.2 mg QE/g DM, and 159.8 mg TE/g DM, respectively. These findings indicate that onion skin extracts could represent a green and promising source of antioxidant phytochemicals. Full article

Background: Thyroid nodules are common, yet only a small proportion are malignant. The independent role of nodule size in malignancy risk remains debated, particularly after adjustment for clinical, biochemical, and sonographic features. Methods: A retrospective cohort study was conducted on adult patients with thyroid nodules evaluated between 2018 and 2025 at a tertiary care center. Clinical, laboratory, ultrasound, cytology, and histopathology data were extracted. Thyroid-stimulating hormone (TSH), free thyroxine (free T4), and sonographic characteristics were analyzed. Univariable and multivariable logistic regression were performed. Missing ultrasound data were addressed using multiple imputation (m = 20), with pooled estimates derived using Rubin’s rules. The final multivariable model included 446 patients. Results: A total of 446 patients were included, of whom 91 (20.4%) had thyroid malignancy. Malignant nodules were significantly larger than benign nodules (2.30 [1.80] cm vs. 1.80 [1.13] cm; p = 0.015). In univariable analysis, TSH, free T4, and multiple ultrasound features were associated with malignancy. In multivariable analysis, nodule size remained the strongest independent predictor of malignancy (adjusted odds ratio [aOR] 1.51; p < 0.001). Hypoechogenicity (aOR 2.07; p = 0.020) and microcalcifications (aOR 1.86; p = 0.047) also remained independently significant, whereas thyroid function parameters were not associated with malignancy after adjustment. Conclusions: Thyroid nodule size is the strongest independent predictor of malignancy, with select ultrasound features retaining additional predictive value. These findings support incorporating nodule size more prominently into thyroid cancer risk stratification while maintaining key sonographic features. Full article