Classification of risk areas using a bootstrap-aggregated ensemble approach for reducing Zika virus infection in pregnant women

Abstract

The Zika (ZIKV) virus has been a potential cause for the birth of children with microcephaly. This situation is unprecedented all over the world, presenting a small number of records in the medical literature. The latest ZIKV epidemic confirms the potential risk of international propagation of mosquito-borne diseases, especially the Aedes albopictus and aegypti species. Therefore, a strategy that aims to define priority areas for intervention is critical to contributing to the prevention and control of these diseases. Hence, this paper proposes the use of an ensemble learning technique, named bootstrap-aggregated ensemble of fine decision trees, to identify epidemic risk areas, reducing its impact on the population, principally in pregnant women. The study analyzed reported cases of ZIKV (n = 112) in Recife, Brazil, during 2015 and 2016. The identification of mosquito activity areas can guarantee a healthy formation for the fetus during and after the gestation through implementation strategies of public policies and further study to control the mosquito vector responsible for these diseases.