Predictions

The image below shows the prediction of the model over the area. If a species is predicted to be present the area is green, if absent red. Confusion tables and total accuracy as determined on the training set and an independent test set follows.

absent

present

VERIFY - predictive accuracy of rules using training data Confusion Matrix: Actual Present Absent Predict Present 0.504 0.066 Absent 0.000 0.429 Conflicts 0 0 Unpredicted 0 0 Predicted 1.000 Accuracy 0.934 s.d. 0.005 Overall Acc. 0.934 s.d. 0.005 VERIFY - predictive accuracy of rules using training data Confusion Matrix: Actual Present Absent Predict Present 0.500 0.072 Absent 0.000 0.428 Conflicts 0 0 Unpredicted 0 0 Predicted 1.000 Accuracy 0.928 s.d. 0.005 Overall Acc. 0.928 s.d. 0.005 PREDICT - use model to produce images