StatEcoNet: Statistical Ecology Neural Networks for Species Distribution Modeling

This paper focuses on a core task in computational sustainability and statistical ecology: species distribution modeling (SDM). In SDM, the occurrence pattern of landscape is predicted by environmental features based observations at set locations. At first, SDM may appear to be binary classification problem, one might inclined employ classic tools (e.g., logistic regression, support vector machines, neural networks) tackle it. However, wildlife surveys introduce structured noise (especially under-counting) observations. If unaccounted for, these observation errors systematically bias SDMs. To address unique challenges this proposes framework called StatEcoNet. Specifically, work employs graphical generative model ecology serve as skeleton proposed carefully integrates networks under framework. The advantages StatEcoNet over related approaches are demonstrated simulated datasets well bird data. Since SDMs critical for ecological science natural resource management, offer boosted analytical powers wide range applications that have significant social impacts, e.g., study conservation threatened species.