Table 3.
Summary of the results obtained by the selected modeling approach (i.e., ANN with 9 neurones in the hidden layer) on both validation and test conditions.
DataSet | Accuracy | F-Score | Sensitivity | Specificity | ||||
---|---|---|---|---|---|---|---|---|
Val. | Test | Val. | Test | Val. | Test | Val. | Test | |
CH | 99,0% | 97,9% | 99,0% | 97,0% | 99,9% | 97,5% | 98,6% | 98,3% |
SBN1E-6 | 90,4% | 85,7% | 90,6% | 86,2% | 90,5% | 85,4% | 90,9% | 86,3% |
NB50 | 91,4% | 88,2% | 91,7% | 88.5% | 91,1% | 88,6% | 92.1% | 87.7% |
NBN50 | 92,4% | 89,8% | 92,5% | 90,1% | 93,6% | 90.7% | 91,3% | 88.8% |