Table 4.
Comparison of the performance of different methods on training data using 20 selected features
| Algorithms | Performance measurement |
|||||
|---|---|---|---|---|---|---|
| ACC | Sn | Sp | PPV | FPR | AUC | |
| K-nearest neighbors (kNN) | 95.511 | 83.3 | 99.2 | 96.7 | 0.8 | 0.966 |
| Support vector machine (SVM) | 95.528 | 85.1 | 98.6 | 94.8 | 1.4 | 0.972 |
| Artificial neural network (MLP) | 95.283 | 86.5 | 97.9 | 92.4 | 2.1 | 0.964 |
| Decision tree (J48) | 94.581 | 84.4 | 97.6 | 91.3 | 2.4 | 0.920 |
| RBF networks (RBFNets) | 94.352 | 86.4 | 96.7 | 88.7 | 3.3 | 0.968 |
| Rule based (RIPPER) | 94.809 | 84.0 | 98.0 | 92.6 | 2.0 | 0.923 |
| Naïve bayes (NB) | 93.585 | 85.5 | 96.0 | 86.4 | 4.0 | 0.955 |
| Random forest (RF) | 95.283 | 86.7 | 97.8 | 92.1 | 2.2 | 0.965 |
Sn = Sensitivity, Sp = Specificity, PPV = Positive predictive value, ACC = Accuracy, FPR = False positive rate and AUC = Area under ROC curve. The highest values are in bold.