Resampling techniques applied to data analysis of biodiversity in Atlantic Forest

Research conducted by Universidade Estadual Paulista Julio de Mesquita Filho (UNESP, Rio Claro, Brazil, and Texas Tech University, project funded by FAPESP and CNPq) involving ecological modelling to understand multiscalebat community responses to landscape degradation. We built non-linear mixed models and resampled it 10,000 times using bootstrap technique in order to get model selection frequencies and best evaluation of null models. This wasvery time consuming and was achieved using Microsoft Azure machines, important tools to conclude tasks underlying large datasets.

More information: