P. Caplat

Weak interactions cause poor performance of common network inference models

The paper evaluates four network inference methods and finds they poorly distinguish interactions in networks with weak interactions and limited data.

Modelling the distribution of rare invertebrates by correcting class imbalance and spatial bias

This paper improves species distribution models for rare millipedes by addressing class imbalance and spatial bias through spatial under-sampling, showing increased sensitivity for rare species.

Data quantity is more important than its spatial bias for predictive species distribution modelling

This study shows that sample size and modelling method are more critical than spatial bias for species distribution model performance.