Table 5. Confusion matrix of plant types, based on neural network model prediction.
Predicted groups | Observed groups | ||
Gypsovag | Narrow endemism | Gypsophile | |
Gypsovag | 15 | 2 | 0 |
Narrow endemism | 3 | 9 | 0 |
Gypsophile | 0 | 0 | 15 |
Model fit | |||
MAE | 0.211 | 0.196 | 0.068 |
RMSE | 0.305 | 0.287 | 0.098 |
Values give the number of each type given in the left-hand column that are categorised as each type given in the top row. Bottom rows indicate mean absolute error (MAE) and RMSE (Root Mean Squared Error) values for each output node in the neural network models, in relation to the output errors (range [0, 1]).