| Methodology adopted | Accuracy (%) | Sensitivity (%) | Specificity (%) | Selected features |
| Evolutionary sigmoidal unit neural network (ESUNN) [21] | 83.22 | 84.32 | 81.65 | 3, 8, 9, 11, 12 | Evolutionary product unit neural network (EPUNN) [21] | 81.89 | 83.67 | 84.91 | 8, 9, 11, 12 | Multilogistic regression + EPUNN [22] | 83.12 | 78.15 | 80.59 | 8, 9, 11, 12, 13 | Mean selection method [16] | 84.44 | 85 | 84 | 3, 8, 9, 11, 12, 13 | Half selection method [16] | 84.81 | 85 | 84 | 3, 8, 9, 10, 11, 12, 13 | Neural network for threshold selection [16] | 85.19 | 85 | 86 | 3, 11, 12, 13 | PSO + ELM | 85.88 | 86.00 | 86.03 | 3, 11, 12, 13 | Proposed SRLPSO + ELM | 89.96 | 87.79 | 88.42 | 11, 12, 13 |
|
|