Table 4. Significance of effect of training area, size of training dataset, and set of bioclimatic variables on AUC values.
Model | Area 1980/1955 | Input variables | Size | |||||||
6/19var | PCA/19var | PCA/6var | 20/10 | Big/Small | ||||||
1955 | 1980 | 1955 | 1980 | 1955 | 1980 | 1955 | 1980 | |||
BIOCLIM | *** | * | *** | ** | *** | ** | ** | NS | NS | *** |
CSM | *** | NS | NS | NS | NS | NS | NS | NS | NS | NS |
DKGARP | . | *** | *** | *** | *** | ** | *** | NS | NS | *** |
DOMAIN | *** | NS | NS | NS | *** | NS | *** | NS | NS | NS |
ED | *** | NS | NS | NS | *** | NS | *** | NS | NS | NS |
ES | NS | NS | NS | NS | NS | NS | NS | NS | NS | *** |
MAXENT | *** | NS | NS | *** | *** | *** | *** | NS | NS | NS |
OMGARP | . | *** | *** | *** | *** | NS | . | NS | NS | NS |
SVM | *** | NS | NS | NS | * | . | * | NS | NS | NS |
Area 1980 vs. 1955, 6 variables vs. 19 variables, first three principal components (PCA) vs. 19 variables, PCA vs. 6 variables, training dataset size = 20 vs. 10, big training dataset (more than 50 presence points) vs. small (less than 50).
***p<0.001 |
**p<0.01 |
*p<0.05 |. p<0.1 | NS not significant.