Practice III: Evaluating predictive performance and setting decision thresholds

This is the fifth part of a training course on Species Distribution Modelling (also called Ecological Niche Modelling) taught by Richard Pearson at University College London in November 2014. The talk describes strategies for generating data to test a model, ways to set a threshold to give binary (present/absent or suitable/unsuitable) predictions, and details two of the most common evaluation statistics: success rate combined with binomial significance and Area Under the ROC curve (AUC). Funding was provided by NERC (the UK Natural Environment Research Council).

[@ryanlam986]

Around the 14 minute mark, you referenced an author by the name of "Liu" as additional literature to read about presence and absence data. Can you clarify his full name and potentially a publication that you would recommend?

[@UCLCBER]

Hi Ryan, sure it can be found here: Canran Liu et al. 2005. Selecting thresholds of occurrence in the prediction of species distributions. Ecography 28:385-393.

[@belayejigu6424]

Liu et al., 2005

[@qasimahmad6714]

Hi..i am absolutely beginner.... i have only presence data... but what i am watching in different videos , people are taking background data for MaxEnt software? from where it come. Do Maxent automatically generate this background file. If not then how can i make background data file especially in MaxEnt software.....? is it possible to separately make file in R and then used that file for MaxEnt software..