Presence-only species distribution models are sensitive to sample prevalence: Evaluating models using spatial prediction stability and accuracy metrics
Published in Ecological Modelling, 2020
Using virtual species we test how spatial predictions of presence-background species distribution models vary with sample size. Of the four algorithms tested, Maxent was the most stable with spatial predictions varying much less in comparison to the algorithms. Machine learning methods Random Forest and Support Vector Machines are sensitive to background sample size, with significant differences in spatial predictions observed.
Recommended citation: Grimmett, L., Whitsed, R., & Horta, A. (2020). Presence-only species distribution models are sensitive to sample prevalence: Evaluating models using spatial prediction stability and accuracy metrics. Ecological Modelling, 431(June), 109194. https://doi.org/10.1016/j.ecolmodel.2020.109194