Table 3.
Comparison of the confusion matrix values and performance for six classifiers (testing data).
| TH | TP | FP | FN | TN | SN (%) | SP | PP | ACC | AUC | |
|---|---|---|---|---|---|---|---|---|---|---|
| RFC | 0.077 | 223 | 960 | 147 | 3804 | 60.27 | 79.85 | 18.85 | 78.44 | 77.59 |
| ADB | 0.487 | 234 | 928 | 136 | 3836 | 63.24 | 80.52 | 20.14 | 79.28 | 79.25 |
| GNB | 0.065 | 258 | 1396 | 112 | 3368 | 69.73 | 70.7 | 15.6 | 70.63 | 78.08 |
| KNNC | 0.065 | 219 | 892 | 151 | 3872 | 59.19 | 81.28 | 19.71 | 79.68 | 72.11 |
| SVC | 0.065 | 221 | 1380 | 149 | 3384 | 59.73 | 71.03 | 13.8 | 70.22 | 71.51 |
| DNN | 0.13 | 238 | 688 | 132 | 4076 | 64.32 | 85.56 | 25.7 | 84.03 | 83.48 |
TH, threshold; RFC, random forest classifier; ADB, AdaBoost classifier: GNB, Gaussian naive Bayes; KNNC, K-nearest neighbor classifier; SVC, support vector machine; DNN, deep neural network.