Table 2 Optimal discriminate analysis models for R. aurora occurrence addressing the seven variables examined
| Â | Training analysisa | LOO analysisb | ||||||
|---|---|---|---|---|---|---|---|---|
Variable | Model | Number | Accuracy (%) | p< | ESS | Accuracy (%) | p< | ESS |
AREA | – | – | – | 0.9696 | 19.1 | – | – | – |
BULL | – | – | – | 0.9999 | 4.8 | – | – | – |
CON | If CON ≤ 96.26 predict no occurrence | 7 | 85.7 | 0.0009 | 81.0 | 71.4 | 0.0039 | 61.9 |
|  | If CON > 96.26 predict occurrence | 21 | 95.2 |  |  | 90.5 | – | – |
EMG | – | – | – | 0.0921 | 52.4 | – | – | – |
FISH | – | – | – | 0.0615 | 42.9 | – | – | – |
FRP | If FRP ≤ 50.1% predict no occurrence | 7 | 100.0 | 0.0238 | 61.9 | – |  |  |
|  | If FRP > 50.1% predict occurrence | 21 | 61.9 |  |  |  | 0.0604 | 42.9 |
PCF | If PCF ≤ 28.0% predict no occurrence | 7 | 85.7 | 0.0025 | 76.2 | 71.4 | 0.0095 | 57.1 |
|  | If PCF > 28.0% predict occurrence | 21 | 90.5 |  |  | 85.7 |  |  |