83 research outputs found
Multitemporal Spectral Analysis for Algae Detection in an Eutrophic Lake using Sentinel 2 Images
No full textEutrophication is characterized by excessive plant and algal growth due to the increased of organic matter, carbon dioxide and nutrients in water body. Although eutrophication naturally occurs over centuries as lakes age, human activities have accelerated it processes and caused dramatic changes to the aquatic ecosystems including elevated algae blooms and risk for hypoxia as well as degradation in the quality of drinking water and fisheries. Monitoring eutrophic processes is therefore highly important to human health and to the aquatic environment. However, the spatial and seasonal distribution of the phenomena and its dynamic are difficult to be resolved using conventional methods as water sampling or sparse acquisition of remote sensing data. This research work proposes a methodology that takes advantage of the high temporal resolution of Sentinel-2 (S2) for monitoring eutrophic reservoir. Specifically, it uses large temporal series of S2 images and advanced temporal unmixing model to estimate the abundance of [Chl-a] and algae species in San Roque reservoir, Argentina, in the period August 2016 to August 2019. The spatial patterns and the temporal tendencies of these aquatic indicators, that have a direct link to Eutrophication, were analysed and evaluated using in situ data in order to assess their contribution to the local water management.Fil: German, Alba. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba; ArgentinaFil: Ferral, Anabella. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba; ArgentinaFil: Scavuzzo, Carlos Matias. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba; ArgentinaFil: Shimoni, M.. Belgian Royal Military Academy; Bélgic
Intestinal parasitic infections in a community from Pampa del Indio, Chaco (Argentina) and their association with socioeconomic and environmental factors
No full textNeglected tropical diseases are a group of 20 disabling diseases, which, in particular, are the most common chronic infections in the most vulnerable people. This study aimed to characterize the infection by intestinal parasites (IPs) in dwellings from a peri-urban neighborhood in Pampa del Indio, Chaco (Argentina), and its association with socioeconomic and environmental variables. Single stool samples were collected from all individuals older than 1 year through household visits and processed using coprological sedimentation and flotation techniques. Standardized questionnaires were used at the household level to collect socio-economic information. Environmental variables were obtained from the Planetscope image, Landsat 8 images and remote sensors, while land-use layers were obtained through the use of a maximum likelihood algorithm. Stool samples were provided by 314 individuals. The prevalence of IPs found was 30.6% (n = 96), with a predominance of Giardia lamblia (12.7%, n = 40) and Hymenolepis nana (7.6%, n = 24). The only soil-transmitted helminth found was Strongyloides stercoralis with a 2.5% prevalence (n = 8). Individuals of adult age (> 18 years) were 0.65 times less likely to present parasitic infections with respect to children and adolescents. The only environmental variable that was closely associated with the presence of IPs, was the Normalized Difference Water Index (NDWI), a measure of humidity; being higher around houses with positive individuals. Most of the IPs found in this study were of water-borne transmission and those transmitted directly from person-to-person, therefore fecal contamination is present. We believe that the low prevalence of STH in this area, which requires a passage through the soil, is related to the environmental characteristics, which are unsuitable for the development/permanence of the infective stages of these parasites. The geospatial data and tools used herein proved to be useful for the study of the relationship between the different factors that influence the presence of IPs in a community, from an eco-health approach.Fil: Scavuzzo, Carlos Matias. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Fundación Mundo Sano; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Delgado, Cintia. Fundación Mundo Sano; ArgentinaFil: Goy, Marcia. Hospital Dr. Dante Tardelli; ArgentinaFil: Crudo, Favio. Fundación Mundo Sano; ArgentinaFil: Porcasi, Ximena. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; ArgentinaFil: Periago, Maria Victoria. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Fundación Mundo Sano; Argentin
La diversificación productiva y de destinos de comercialización mejoran los precios finales de venta en horticultura periurbana
No full textLos sistemas hortícolas periurbanos están en permanente adaptación para compatibilizar la actividad productiva con la vecindad al territorio habitado y con las demandas de mercados y consumidores. Su rentabilidad depende en gran medida de los altamente dinámicos precios diarios de venta. El objetivo del trabajo fue evaluar cómo la diversificación de especies hortícolas y de prácticas productivas pueden potenciar mercados, y mejorar los precios finales de venta en quintas hortícolas periurbanas de la Ciudad de Córdoba. Para ello se realizaron encuestas y muestreos mensuales en 20 quintas representativas durante el periodo 09/2018-12/2020. Se relevó información sobre el número de especies hortícolas presentes en cada quinta, la diversidad de prácticas productivas (asociación y rotación de cultivos, producción de semillas, uso de bioinsumos/agroquímicos, abonos orgánicos, MIP, mulching natural, labranza mínima, cultivo de cobertura), y los destinos y precios de venta. Los datos se analizaron con métodos multivariados (Análisis de Conglomerados y Coordenadas Principales) con el programa Infostat y se construyó un “índice de precios hortícolas” (IP). Como resultado se observó una correlación positiva entre mayor valor del IP y las quintas con mayor diversificación de especies hortícolas, prácticas productivas y destinos de venta. Los mayores precios de venta los obtuvieron los productores que no aplican agroquímicos, comercializan directo a consumidores, en ferias o bolsones. De los 20 casos estudiados, las cuatro quintas que comercializan en la Feria Agroecológica local se vincularon con los mejores precios de venta, donde la diversificación de las prácticas productivas es un criterio de importancia en el Sistema Participativo de Garantía.Fil: Marinelli, María Victoria. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Argüello Caro, Evangelina Beatriz. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Universidad Nacional de Córdoba. Facultad de Ciencias Agropecuarias. Cátedra de Fisiología Vegetal; Argentina. Instituto Nacional de Tecnología Agropecuaria. Centro de Investigaciones Agropecuarias. Instituto de Patología Vegetal; ArgentinaFil: Giobellina, Beatriz. Instituto Nacional de Tecnología Agropecuaria. Centro Regional Córdoba; ArgentinaFil: Giannini Kurina, Franca. Instituto Nacional de Tecnologia Agropecuaria. Centro de Investigaciones Agropecuarias. Unidad de Fitopatologia y Modelizacion Agricola. Grupo Vinculado Catedra de Estadistica y Biometria de la Facultad de Ciencias Agropecuarias de la Universidad Nacional de Cordoba Al Ufyma | Consejo Nacional de Investigaciones Cientificas y Tecnicas. Centro Cientifico Tecnologico Conicet - Cordoba. Unidad de Fitopatologia y Modelizacion Agricola. Grupo Vinculado Catedra de Estadistica y Biometria de la Facultad de Ciencias Agropecuarias de la Universidad Nacional de Cordoba Al Ufyma.; ArgentinaFil: Scavuzzo, Carlos Matias. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina41º Congreso Argentino de Horticultura; V Simposio de Aromáticas, Medicinales y CondimenticiasLa PlataArgentinaAsociación Argentina de Horticultur
Explorando el uso de herramientas de sensores remotos y tecnologías geoespaciales aplicadas al problema multidimensional de la seguridad alimentaria
No full textEl objetivo de este estudio fue analizar qué papel pueden jugar las tecnologías de teledetección para estudiar los factores multidimensionales que influyen en la seguridad alimentaria y nutricional (SAN), en Córdoba Argentina. [Metodología] El área de estudio comprende la ciudad de Córdoba, Argentina. Se obtuvieron datos epidemiológicos de la prevalencia de bajo peso, sobrepeso y obesidad (malnutrición) durante el 2013 de 23 centros de atención primaria de la ciudad. Se exploraron las condiciones ambientales de los centros en un radio de 1000 m. Se clasificaron imágenes SPOT 5, se utilizaron características espectrales y espaciales y se evidencia cómo una clasificación no supervisada puede dar información para describir la dimensión social y el acceso económico a los alimentos. Se realizó una regresión lineal multivariante para examinar la relación entre la prevalencia de malnutrición y las variables ambientales y espaciales, derivadas de las imágenes SPOT. [Resultados] Los resultados de la clasificación no supervisada de imágenes muestran la diferencia en el patrón espectral-espacial de los barrios, evidencian cómo una simple clasificación de imágenes de satélite puede convertirse en una herramienta de discriminación útil. Se obtienen análisis de regresión múltiple con R2 ajustados de 0,70 y 0,6435 respectivamente para desnutrición, y sobrepeso y obesidad. A partir de los modelos obtenidos, se construyen mapas continuos de prevalencia. [Conclusiones] El método propuesto en este trabajo puede discriminar socialmente diferentes áreas relacionadas con la SAN. Es innovador y necesario aprovechar las herramientas de teledetección y las tecnologías geoespaciales, en nuestra región, aplicadas a la SAN.Di Fino, E., Scavuzzo, C., Campero, M., Scavuzzo, C., & Defagó, M. (2022). Exploring the use of remote sensing tools and geospatial technologies applied to the multidimensional food security problem. Uniciencia, 36(1), 1-15. https://doi.org/10.15359/ru.36-1.4
Implementation of a proactive system to monitor Aedes aegypti populations using open access historical and forecasted meteorological data
No full textDue to the global increase in mosquito-borne diseases outbreaks it is recommended to increase surveillance and monitoring of vector species to respond swiftly and with early warning indicators. Usually, however, the information about vector presence and activity seems to be insufficient to implement timely and effective control strategies. Here we present an improved mathematical model of Aedes aegypti population dynamics with the aim of making the Dengue surveillance system more proactive. The model considers the four life stages of the mosquito: egg, larva, pupa and adult. As driving factors, it incorporates temperature which affects development and mortality rates at certain stages, and precipitation which is known to affect egg submergence and hatching, as well as larval mortality associated with desiccation. Our mechanistic model is implemented as a free and stand-alone system that automatically retrieves all needed inputs, runs a simulation and shows the results. A major improvement in our implementation is the capacity of the system to predict the population dynamics of Ae. aegypti in the near future, given that it uses gridded weather forecast data. Hence, it is independent by meteorological station proximity. The model predictions are compared with field data from Córdoba City, Argentina. Although field data have high variability, an overall accordance has been observed. The comparison of results obtained using observed weather data, with the simulations based on forecasts, suggests that the modeled dynamics are accurate up to 15 days in advance. Preliminary results of Ae. aegypti population dynamics for a consecutive three-year period, spanning different eco-regions of Argentina, are presented, and demonstrate the flexibility of the system.Fil: Aguirre, Exequiel. Comision Nacional de Actividades Espaciales; Argentina. Universidad Nacional de Córdoba. Facultad de Matemática, Astronomía y Física; ArgentinaFil: Andreo, Verónica Carolina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Porcasi Gomez, Ximena. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Lopez, Laura. Provincia de Córdoba. Ministerio de Salud; ArgentinaFil: Guzman, Claudio. Provincia de Córdoba. Ministerio de Salud; ArgentinaFil: González, Patricia. Provincia de Córdoba. Ministerio de Salud; ArgentinaFil: Scavuzzo, Carlos Matias. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentin
Quality and environmental footprints of diets by socio-economic status in Argentina
No full textFood consumption and production plays a major role on human health and environmental sustainability. In addition, socio-economic status (SES) could be an important determinant on these impacts. Here we investigated the quality and environmental footprints of diets of different SES in Argentina. Dietary data was obtained from a recent national-wide representative survey (21,547 households), and individual intake was estimated by applying the Adult-Male Equivalent (AME) approach. Dietary quality was assessed by comparing the apparent consumption of foods and nutrients with the level of intake associated with the lowest risk of mortality, and applying the Alternate Healthy Eating Index 2010 (AHEI-2010). The environmental footprints were assessed with a set of six environmental indicators: greenhouse gas (GHG) emissions, total land occupation, cropland demand, fossil energy use, freshwater consumption and eutrophying emissions. Per capita total expenditure was utilized as an indicator of SES. We found a suboptimal intake of healthy and excess of unhealthy foods and nutrients in all SES, as well as a low AHEI-2010 score (34.7 ± 3.56 over 100). Regarding environmental footprints per AME, we estimated 8.91 ± 2.49 kg CO2-eq/day for GHG emissions, 9.30 ± 2.93 MJ/day for fossil energy use, 54.2 ± 14.8 m2/day for total land occupation, 7.07 ± 1.90 m2/day for cropland demand, 205 ± 65.3 L/day for freshwater consumption and 19.8 ± 6.43 g PO4-eq/day for eutrophying emissions. The Argentinian diet was found to be both unhealthy and unsustainable in all socio-economic levels.Fil: Arrieta, Ezequiel Martín. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba. Instituto Multidisciplinario de Biología Vegetal. Universidad Nacional de Córdoba. Facultad de Ciencias Exactas Físicas y Naturales. Instituto Multidisciplinario de Biología Vegetal; ArgentinaFil: Geri, Milva. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Bahía Blanca. Instituto de Investigaciones Económicas y Sociales del Sur. Universidad Nacional del Sur. Departamento de Economía. Instituto de Investigaciones Económicas y Sociales del Sur; ArgentinaFil: Becaria Coquet, Julia. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba; Argentina. Universidad Nacional de Córdoba. Facultad de Medicina. Escuela de Nutrición; ArgentinaFil: Scavuzzo, Carlos Matias. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Córdoba; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Universidad Nacional de Córdoba; Argentina. Fundación Mundo Sano; ArgentinaFil: Zapata, Maria Elisa. Centro de Estudio S/nutrición Infantil; ArgentinaFil: Gonzalez, Alejandro Daniel. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Patagonia Norte. Instituto Andino Patagónico de Tecnologías Biológicas y Geoambientales. Universidad Nacional del Comahue. Instituto Andino Patagónico de Tecnologías Biológicas y Geoambientales; Argentina. Universidad Nacional del Comahue. Centro Regional Universitario Bariloche; Argentin
Spatial patterns of intestinal parasite infections among children and adolescents in some indigenous communities in Argentina
No full textArgentina has a heterogeneous prevalence of infections by intestinal parasites (IPs), with the north in the endemic area, especially for soil-transmitted helminths (STHs). We analyzed the spatial patterns of these infections in the city of Tartagal, Salta province, by an observational, correlational, and cross-sectional study in children and adolescents aged 1 to 15 years from native communities. One fecal sample per individual was collected to detect IPs using various diagnostic techniques: Telemann sedimentation, Baermann culture, and Kato-Katz. Moran’s global and local indices were applied together with SaTScan to assess the spatial distribution, with a focus on cluster detection. The extreme gradient boosting (XGBoost) machine-learning model was used to predict the presence of IPs and their transmission pathways. Based on the analysis of 572 fecal samples, a prevalence of 78.3% was found. The most frequent parasite was Giardia lamblia (30.9%). High- and low-risk clusters were observed for most species, distributed in an east-west direction and polarized in two large foci, one near the city of Tartagal and the other in the km 6 community. Spatial XGBoost models were obtained based on distances with a minimum median accuracy of 0.69. Different spatial patterns reflecting the mechanisms of transmission were noted. The distribution of the majority of the parasites studied was aligned in a westerly direction close to the city, but the STH presence was higher in the km 6 community, toward the east. The purely spatial analysis provides a different and complementary overview for the detection of vulnerable hotspots and strategic intervention. Machine-learning models based on spatial variables explain a large percentage of the variability of the IPs.Fil: Scavuzzo, Carlos Matias. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Universidad Nacional de Córdoba; ArgentinaFil: Campero, Micaela Natalia. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Universidad Nacional de Córdoba; ArgentinaFil: Maidana, Rosana Elizabeth. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Universidad Nacional de Córdoba; ArgentinaFil: Oberto, María Georgina. Universidad Nacional de Córdoba; ArgentinaFil: Periago, Maria Victoria. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Fundación Mundo Sano; ArgentinaFil: Porcasi Gomez, Ximena. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentin
Towards a workflow for operational mapping of Aedes aegypti at urban scale based on remote sensing
No full textRemote sensing (RS) applications for vector borne diseases are a field of high social impact increasingly relevant in the context of a higher frequency of Dengue, Chikungunya and Zika outbreaks at global scale and especially in Latin America. The operative use of RS technologies is however still rare. Therefore, the objective of this work is to generate and analyze multitemporal Aedes aegypti's suitability maps and to share the open source tools used towards the building of an operative workflow. As a proof of concept, we implemented a process chain to obtain maps for Ae. aegypti activity within the 2017–2018 mosquito breeding season based on ovitraps records and RS data within the framework of ecological niche modeling. The workflow was carefully thought as to consider possible biases in training data, model calibration to attain the best hyper-parameter combination, model selection, variable selection and validation with independent data. The predictive maps showed high suitability for Ae. aegypti within the city, except in large vegetated areas and the commercial downtown consistently with previous studies and our own observations. Relevant variables included distance to built-up surfaces, distance to vegetated areas and correlation, a texture measure reflecting surface heterogeneity. Validation results suggested that the spatial distribution of ovitraps should be re-examined. All the steps in the proposed workflow were implemented using freely available and open source software, which warrants reproducibility and allows for re-use and modifications in terms of methods and RS or mosquito data available.Fil: Andreo, Verónica Carolina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Cuervo Bustamante, Pablo Fernando. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Santa Fe. Instituto de Ciencias Veterinarias del Litoral. Universidad Nacional del Litoral. Facultad de Ciencias Veterinarias. Instituto de Ciencias Veterinarias del Litoral; ArgentinaFil: Porcasi Gomez, Ximena. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Lopez, Laura. No especifíca;Fil: Guzman, Claudio. No especifíca;Fil: Scavuzzo, Carlos Matias. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentin
Intestinal parasites in children from native communities of Salta, Argentina
No full textIntroduction: Intestinal parasites particularly affect vulnerable populations, making their management crucial in endemic areas. Objective: To evaluate the relationship between intestinal parasites, socioeconomic characteristics, and anthropometric nutritional status. Methods: This is an observational, cross-sectional study conducted on children aged 1 to 15 years in Tartagal, Argentina, for a year. Intestinal parasitic infections were determined through fecal analysis using sedimentation techniques (Teleman), Kato-Katz, and the Baerman method. Demographic and socioeconomic data were collected using a standardized survey. Height, weight, and mid-upper arm and head circumference were measured in the participants. Results: Of the 588 participants who provided fecal samples, 55.4% had intestinal parasites, with G. lamblia and H. nana being the most prevalent species. Additionally, 31.9% of participants were found to have malnutrition due to excess weight. Risk factors for intestinal parasite presence included age, overcrowding, and walking barefoot, while owning a refrigerator, having an animal corral, adequate water storage, and belonging to the Guaraní ethnic group were protective factors. High vulnerability areas were observed regarding housing conditions and access to public services. Conclusions: An association between socioeconomic conditions and the risk of intestinal parasite presence was observed. However, no association was found between anthropometric nutritional status and the presence of intestinal parasites.Introducción: Los parásitos intestinales afectan particularmente a grupos poblacionales vulnerables siendo de importancia su abordaje en zonas endémicas. Objetivo: Evaluar la relación de parásitos intestinales con características socioeconómicas y estado nutricional antropométrico. Métodos: Estudio observacional, y transversal en niños de 1 a 15 años de Tartagal, Argentina durante un año. Las parasitosis intestinales se determinaron por análisis de materia fecal usando técnicas de sedimentación (Teleman) y Kato Katz, y el método de Baerman. Los datos demográficos y socioeconómicos se colectaron con una encuesta estandarizada. Se midió talla, peso, perímetro braquial y cefálico a los participantes. Resultados: De los 588 participantes que entregaron materia fecal, el 55.4% presentó parásitos intestinales, siendo Giardia lamblia e Hymenolepis nana las especies más prevalentes. El 31.9% de los participantes presentaron malnutrición por exceso. Los factores de riesgo para la presencia de parásitos intestinales fueron la edad, el hacinamiento y caminar descalzo, mientras que tenencia de heladera, corral para animales, almacenamiento de agua adecuado y pertenecer a la etnia Guaraní, actuaron como factores protectores. Se observaron zonas de alta vulnerabilidad de las viviendas y de acceso a servicios públicos. Conclusiones: Se evidenció una asociación entre las condiciones socioeconómicas y el riesgo de presencia de parásitos intestinales. Sin embargo, no se observó asociación entre el estado nutricional antropométrico y la presencia de parásitos intestinales.Fil: Scavuzzo, Carlos Matias. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; ArgentinaFil: Campero, Micaela Natalia. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; ArgentinaFil: Oberto, María Georgina. Universidad Nacional de Córdoba; ArgentinaFil: Porcasi Gomez, Ximena. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Comision Nacional de Actividades Espaciales. Instituto de Altos Estudios Espaciales "Mario Gulich"; ArgentinaFil: Periago, Maria Victoria. Fundación Mundo Sano; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentin
Feature importance: Opening a soil-transmitted helminth machine learning model via SHAP
No full textIn the field of landscape epidemiology, the contribution of machine learning (ML) to modeling of epidemiological risk scenarios presents itself as a good alternative. This study aims to break with the "black box" paradigm that underlies the application of automatic learning techniques by using SHAP to determine the contribution of each variable in ML models applied to geospatial health, using the prevalence of hookworms, intestinal parasites, in Ethiopia, where they are widely distributed; the country bears the third-highest burden of hookworm in Sub-Saharan Africa. XGBoost software was used, a very popular ML model, to fit and analyze the data. The Python SHAP library was used to understand the importance in the trained model, of the variables for predictions. The description of the contribution of these variables on a particular prediction was obtained, using different types of plot methods. The results show that the ML models are superior to the classical statistical models; not only demonstrating similar results but also explaining, by using the SHAP package, the influence and interactions between the variables in the generated models. This analysis provides information to help understand the epidemiological problem presented and provides a tool for similar studies.This study was funded by Fundación Mundo Sano and Instituto de Salud Carlos III. The funders had no roles in the design of the study or collection, analysis and interpretation of the data. C.M.S. and M.N.C. had a PhD scholarship from Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET).S
- …
