Search

Electronic health records provide rich, heterogeneous data about the evolution of the patients’ health status. However, such data need to be processed carefully, with the aim of extracting meaningful information for clinical decision support. In this paper, we leverage interpretable (deep) learning and signal processing tools to deal with multivariate time-series data collected from the Intensive Care Unit (ICU) of the University Hospital of Fuenlabrada (Madrid, Spain). The presence of antimicrobial multidrug-resistant (AMR) bacteria is one of the greatest threats to the health system in general and to the ICUs in particular due to the critical health status of the patients therein. Thus, early identification of bacteria at the ICU and early prediction of their antibiotic resistance are key for the patients’ prognosis. While intelligent data-based processing and learning schemes can contribute to this early prediction, their acceptance and deployment in the ICUs require the automatic schemes to be not only accurate but also understandable by clinicians. Accordingly, we have designed trustworthy intelligent models for the early prediction of AMR based on the combination of meaningful feature selection with interpretable recurrent neural networks. These models were created using irregularly sampled clinical measurements, both considering the health status of the patient and the global ICU environment. We explored several strategies to cope with strongly imbalance data, since only a few ICU patients are infected by AMR bacteria. It is worth noting that our approach exhibits a good balance between performance and interpretability, especially when considering the difficulty of the classification task at hand. A multitude of factors are involved in the emergence of AMR (several of them not fully understood), and the records only contain a subset of them. In addition, the limited number of patients, the imbalance between classes, and the irregularity of the data render the

The main dietary challenge for organic swine feeding is ensuring that feed fulfils the nutrient requirements, in particular to protein and amino acid profile, after the expiration that allow producers to include up to 5% non-organic feed within their rations, which is expected to 2020. Thus, this work was designed to propose an organic feeding strategy and to analyse the nutritional, economic and environmental viability of locally produced feedstuff use. A representative sample of organic feed ingredients were selected, according to EC No 834/2007 and No 889/2008, in feed mills and producers from Catalonia and nearby area: cereals (n=22), cereal by-products (n=4) and protein concentrate (n=25). Diets were formulated for the different physiological phases, adjusted to the Spanish (FEDNA, conventional) and French (ITAB, organic) nutritional recommendations. Diets were formulated without soybean expeller or full-fat soybean when using French standards, using instead other protein sources adapted to the Mediterranean climate (Narbonne vetch, beans, peas, and fenugreek) and winter cereals (rye, wheat, barley and triticale). The assessment of environmental impacts (phosphorus demand, non-renewable energy, eutrophication, acidification, climate change and land occupation) of the diets (per unit of produced feed), through life-cycle analysis by means of the database ECOALIM (2016), was higher in piglets and lower in pregnant sows, and much higher when formulating according to FEDNA vs. ITAB standards. In the economic valuation of the diets (purchase or self-production), a business model that integrates agriculture and livestock farming would be more profitable., El principal desafío en alimentación para porcino ecológico es garantizar que los alimentos de la dieta cumplan los requerimientos nutricionales, en particular en proteína y perfil de aminoácidos, al finalizar la moratoria de uso de hasta un 5% de alimentos no ecológicos en las raciones, prevista para 2020. Así, este trabajo se planteó para proponer una estrategia de alimentación y analizar la viabilidad nutricional, económica y ambiental de formular raciones con alimentos alternativos, producidos localmente. Se seleccionó una muestra representativa de productos de cultivos ecológicos, de acuerdo al Reglamento (CE) 834/2007 y 889/2008, de fábricas de piensos y productores de Cataluña y zona próxima: cereales (n=22), subproductos de cereales (n=4) y concentrados de proteína vegetal (n=25). Las dietas se formularon para las diferentes fases fisiológicas, ajustadas a las recomendaciones nutricionales españolas (FEDNA, convencional) y francesas (ITAB, ecológicas). Se formularon dietas prescindiendo de la soja (torta de presión y soja integral) según recomendaciones nutricionales francesas y utilizando fuentes proteicas adaptadas al clima mediterráneo (alverjón, habas, guisantes y alholva) y cereales de invierno (centeno, trigo, cebada y triticale). Los impactos ambientales de las dietas (por unidad de pienso compuesto producido) en cuanto a demanda de fósforo, uso de energía no renovable, eutrofización, acidificación, cambio climático y ocupación de tierra, evaluados mediante el análisis de ciclo de vida de la base de datos francesa ECOALIM (2016), resultan mayores en lechones y menores en cerdas gestantes, y notablemente superiores si se formula según FEDNA vs. ITAB. En la valoración económica de las dietas (compra frente a autoproducción), resultaría más rentable un modelo de negocio con integración agropecuaria.

[EN] This paper proposes an interpolation model for monthly rainfall in large areas of complex orography. It has been implemented in the Iberian Peninsula (continental territories of Spain and Portugal), Balearic and Canary Islands covering a territory of almost 600.000km(2). To do this a data set that comprises a total number of 11,822 monthly precipitation series has been created (11,042 provided by the Spanish Meteorological Agency and 780 provided by the National Water Resources Information System of the Portuguese Water Institute). The data set covers the period from October 1940 until September 2005. The interpolation model has been based on the assumption of two different components on monthly precipitation. The first component reflects local and seasonal characteristics and 24 different mean monthly precipitation maps (12) and SDs maps (12) compose it. It considers the varying influence of physiographic variables such as altitude and orientation. The second precipitation component reflects the synoptic pattern that dominated each month of the series and it is composed by series of anomalies of monthly precipitation (780). Anomalies have been interpolated by means of ordinary kriging once local spatial continuity was assumed. Gridded maps of each variable have been developed at 200m resolution following a hybrid methodology that implements two different interpolation techniques. The first technique applies a regression analysis to derive maps depending on altitude and orientation; the second one is a weighting technique to consider the non-linearity of the precipitation/altitude dependence. Cross validation has been applied to estimate the goodness of both techniques. Results show an average annual precipitation of 655mm/year. Although this figure is only 4% less than the estimate of MAGRAMA (2004), regional and local differences are highlighted when the spatial distribution is considered. The model constitutes a comprehensive implementation considering the a