Ecological insight through machine learning

In a new paper published in Methods in Ecology and Evolution, Jornada LTER scientists, led by Dr. Quiyan Yu, review salient methods in machine learning algorithms and evaluate their effect on successful ecological inference. A number of recommendations emerged, including the removal of spurious (correlated but not functionally important) variables from ML models, and the use of surrogate models for interpretation of functional relationships.

Paper DOI: