This paper presents a novel approach for accurate half-hourly solar irradiance estimation under diverse sky conditions using the Histogram-based Gradient Boosting (HistGB) algorithm. The proposed models are developed from meteorological data collected by an automated weather station in Argentina and are enhanced through a clearness index ( K_t )–based pre-training stratification, which segments the dataset into subsets associated with distinct sky-condition regimes, thereby improving model adaptability. An exploratory data analysis (EDA) is conducted to evaluate the relevance of the available input features for irradiance estimation. Based on this analysis, a dimensionality reduction strategy is applied, yielding a compact input set consisting of only three variables: month of the year, hour of the day, and air temperature. The proposed approach is benchmarked against other tree-based methods, including Extreme Gradient Boosting (XGBoost) and Random Forest (RF) as well as an Artificial Neural Networks (ANN). In addition to standard validation during training, the selected models are rigorously evaluated using an independent dataset collected after the training phase, demonstrating both robustness and practical applicability. The results show high estimation accuracy, with R^2 values of 0.97 and 0.91 for clear-sky and partly cloudy conditions, respectively, during validation, and 0.97 and 0.90 when tested on unseen data. Correspondingly low error levels are obtained, with RMSE values of 34.54 W/m ^2 and 46.34 W/m ^2 during validation, and 34.25 W/m ^2 and 48.29 W/m ^2 on the independent test dataset. Overall, the proposed framework offers a low-complexity and computationally efficient solution for reliable solar irradiance estimation, making it well suited for real-time applications and low-cost solar energy systems operating under diverse environmental conditions.
Resumen La sindemia de COVID-19 ha tenido efectos desiguales en la población a nivel mundial y ha profundizado las inequidades preexistentes en los servicios de salud. Este artículo analiza las experiencias de profesionales de salud pública en pediatría de la Región Sudeste del Conurbano Bonaerense (Argentina), centrado en las tensiones que transformaron la atención, los cambios en el entorno y el bienestar, y las implicaciones para el futuro de la región. Los datos se recolectaron mediante una encuesta retrospectiva en 2022, con la participación de 74 profesionales de pediatría (enfermería, medicina, administración) en efectores locales. Los resultados evidencian una profundización de barreras estructurales para el acceso a la salud, la falta de apoyo institucional para los/as trabajadores/as de salud, el impacto duradero sobre su bienestar y el cambio en sus hábitos cotidianos y preocupaciones. Del análisis emerge el fenómeno denominado “Efecto Agobio”, que se distingue del “burnout” al ser un fenómeno colectivo y no individual. Es un sentimiento generalizado de desánimo con efectos negativos sobre la gestión de la salud pública. Las conclusiones buscan aportar evidencia sobre este fenómeno y contribuir al desarrollo de estrategias sanitarias más sensibles y sostenibles, basadas en los aprendizajes que dejó la sindemia.
El presente artículo es resultado de un proyecto de investigación en proceso, de carácter cualitativo, exploratorio y descriptivo. Mediante las técnicas de la historia oral, la etnografía y la indagación hemerográfica en publicaciones periodísticas, se busca reconstruir la historia de las mujeres raperas del Gran Buenos Aires en las últimas dos décadas del siglo XX. Este escrito se estructura en dos secciones: la primera corresponde al rol de las mujeres durante la constitución de la vieja escuela del hip hop local; la segunda, a la participación de las mujeres en la escena del rap de la década de 1990. El objetivo es analizar cómo fueron representadas en los discursos de la prensa especializada y, transversalmente, el trato desigual que les dieron con respecto a los raperos varones.
Chronic wound management remains a significant clinical challenge, requiring adaptive therapeutic approaches to achieve wound closure that nonetheless frequently prove fruitless. Balancing the initial pro-inflammatory response with debris removal and tissue rebuilding remains elusive in most cases, leading to pain, drastic quality-of-life deterioration, and, eventually, amputation. Meanwhile, patient adherence is an overarching theme. Furthermore, non-surgical alternatives that effectively promote tissue rebuilding are essential for patients seeking to avoid further invasive procedures. We report a patient with a recalcitrant ulcer managed using human amniotic membrane dressing (hAM-pe) and a bovine collagen matrix (BCM) in spatially distinct areas as an intra-patient control. Methodology included clinical monitoring and ad hoc molecular and histological analyses to assess inflammatory markers and tissue architecture. Following 59 days of observation, the superior evolution of the hAM-pe-treated zone led to the clinical decision to extend hAM-pe treatment over the adjacent BCM area, resulting in total wound closure. The hAM-pe-treated site demonstrated accelerated closure and clinical resolution of inflammation without the presence of a granulomatous response. Molecular analysis revealed downregulated pro-inflammatory mediators (IL-1β, TNF-α, CXCL-10) and upregulated markers associated with angiogenesis (VEGF, CD34) and tissue repair (Arginase-1). In this case, the non-surgical hAM-pe treatment was associated with a favorable healing trajectory, characterized by superior inflammation resolution and enhanced tissue organization (collagen type I/III maturation). While these descriptive findings suggest the potential advantages of amniotic membrane dressings in promoting advanced tissue repair, they remain limited to this individual observation. Further research in larger cohorts is required to validate these mechanisms.
Harmful cyanobacterial blooms (HABs) pose serious risks to freshwater ecosystems, drinking water supplies, and public health, highlighting the need for reliable early-warning systems. This study presents a rigorously validated machine learning framework for predicting cyanobacterial alert levels under strongly imbalanced conditions using routinely measured physicochemical variables. Four gradient boosting algorithms were systematically combined with 12 resampling strategies and evaluated within a nested cross-validation framework to ensure unbiased performance assessment. Model evaluation incorporated metrics tailored to imbalanced classification, including recall, F1-score, balanced accuracy (BA), and the Matthews correlation coefficient (MCC), with particular emphasis on the detection of alert events. Results demonstrate that resampling is critical for improving minority-class detection, with SMOTE-based approaches consistently providing the most favorable balance between sensitivity and precision across algorithms. LightGBM combined with SMOTE achieved the highest recall and F1-score, together with strong BA and MCC values and low variability across folds, indicating robust generalization. XGBoost combined with SMOTE exhibited a more balanced precision-recall profile with comparable overall performance but higher variability. SHAP-based interpretability analyses revealed consistent and ecologically meaningful drivers across models, with water temperature, turbidity, and pH emerging as the most influential predictors. By restricting inputs to variables measurable in near real time using low-cost in situ sensors, the proposed framework is designed to support operationally feasible early-warning applications through frequent updates of alert-level predictions within environmental monitoring systems. Overall, the findings highlight the importance of addressing class imbalance, ensuring rigorous validation, and incorporating interpretability to support practical and operationally feasible cyanobacterial early-warning applications.
