Table 6 The performance evaluation of the proposed features based on PyFeat with AB compared with other techniques: five numerical representations, RFF, Pse-in-One2.0, iLearn, and SubFeat with 10-fold cross-validation based on the protein dataset. Significant values are in bold.
From: Predicting Parkinson disease related genes based on PyFeat and gradient boosted decision tree
Metric | ACC (%) | AUC (%) | AUPR (%) | F1-score (%) | MCC | SEN (%) | SPC (%) |
|---|---|---|---|---|---|---|---|
Binary | 53.0 | 57.6 | 58.4 | 53.1 | 0.061 | 54.1 | 51.9 |
Integer | 61.8 | 64.7 | 66.9 | 60.6 | 0.241 | 60.1 | 63.4 |
Real | 57.1 | 62.2 | 64.0 | 57.5 | 0.142 | 58.5 | 55.7 |
Z-curve | 59.0 | 60.9 | 64.7 | 58.1 | 0.182 | 57.4 | 60.7 |
EIIP | 58.4 | 62.3 | 65.5 | 57.7 | 0.171 | 57.4 | 59.6 |
RFF | 63.4 | 66.4 | 66.6 | 64.9 | 0.271 | 68.3 | 58.5 |
Pse-in-One2.0 | 62.3 | 65.3 | 65.3 | 62.9 | 0.246 | 63.7 | 60.9 |
iLearn | 60.7 | 62.9 | 64.1 | 59.9 | 0.215 | 59.3 | 62.0 |
SubFeat | 59.6 | 63.0 | 67.0 | 58.9 | 0.195 | 57.4 | 61.7 |
Proposed features | 79.4 | 84.9 | 86.0 | 78.7 | 0.590 | 76.8 | 82.1 |
