Construction of hazard maps of Hantavirus contagion using Remote Sensing, logistic regression and Artificial Neural Networks: case Araucan\'ia Region, Chile

TL;DR

This paper develops hazard maps for Hantavirus contagion in Araucanía, Chile, by integrating satellite imagery, biological data, and infection cases using logistic regression and neural networks to model contagion risk.

Contribution

It introduces a combined approach using remote sensing and machine learning models to predict Hantavirus risk areas, which is novel for this region.

Findings

Successful creation of hazard maps indicating high-risk zones.

Demonstrated effectiveness of neural networks in modeling contagion risk.

Provided a methodological framework for similar epidemiological studies.

Abstract

In this research, methods and computational results based on statistical analysis and mathematical modelling, data collection in situ in order to make a hazard map of Hanta Virus infection in the region of Araucania, Chile are presented. The development of this work involves several elements such as Landsat satellite images, biological information regarding seropositivity of Hanta Virus and information concerning positive cases of infection detected in the region. All this information has been processed to find a function that models the danger of contagion in the region, through logistic regression analysis and Artificial Neural Networks