Table 3.
Performance comparison between models.
Multiclass red neuronal | Decision forest multiclass | Multiclass logistic regression | Multiclass jungle of decisions | |
---|---|---|---|---|
General accuracy | 0.784 | 0.831 | 0.773 | 0.835 |
Medium accuracy | 0.892 | 0.887 | 0.886 | 0.890 |
Micro-precision averaged | 0.784 | 0.831 | 0.773 | 0.835 |
Macro-accuracy averaged | 0.446 | 0.691 | 0.374 | 0.817 |
Micro-sensitivity averaged | 0.784 | 0.831 | 0.773 | 0.835 |
Macro-sensitivity averaged | NA | 0.617 | NA | 0.537 |
Artificial neural networks matrix | ||||
Current class | 3 | 4 | 5 | |
3 | 0.971 | 0 | 0.029 | |
4 | 0.972 | 0.001 | 0.028 | |
5 | 0.829 | 0 | 0.17 | |
Decision forest matrix | ||||
Current class | 3 | 4 | 5 | |
3 | 0.939 | 0.033 | 0.027 | |
4 | 0.494 | 0.417 | 0.089 | |
5 | 0.406 | 0.1 | 0.494 | |
Logistic regression matrix | ||||
Current class | 3 | 4 | 5 | |
3 | 0.94 | 0.014 | 0.046 | |
4 | 0.799 | 0.121 | 0.08 | |
5 | 0.595 | 0.035 | 0.37 | |
Jungle matrix of decisions | ||||
Current class | 3 | 4 | 5 | |
3 | 0.987 | 0 | 0.013 | |
4 | 0.778 | 0.153 | 0.069 | |
5 | 0.507 | 0.023 | 0.471 |