Figure 6

Jackknife analysis results. (a) The Jackknife of regularized training gain test which predictor (variable) contributes most to the model (b) The Jackknife of test gain uses the test samples (not used in the model) to measure how each predictor (variable) contributes to the average likelihood of presence locality (c) The Jackknife of AUC assesses differences in significance to show that the model performs better than random based on each variable.