Tables
Table 1 The average performance of the models across 10 crossvalidated folds of the training dataset calculated for random and background weighting (BkWt) schemes. The threshold-dependent measures were calculated given a 10 percentile of suitability score at the training occurrence points.