Table 8 Comparison of classifiers in hybrid meta features based on different datasets (\(\%\)).
From: Hybrid framework for membrane protein type prediction based on the PSSM
Dataset | Method | Se | Sp | F-m | Mcc | OA |
|---|---|---|---|---|---|---|
Dataset1 | Xgboost | 69.81 | 98.67 | 70.49 | 69.34 | 93.24 |
KNN | 77.88 | 99.12 | 79.94 | 80.01 | 94.96 | |
LightGBM | 69.90 | 98.71 | 70.53 | 69.41 | 93.31 | |
SVM | 77.54 | 99.10 | 79.45 | 79.52 | 94.76 | |
RF | 77.23 | 99.06 | 72.65 | 72.37 | 94.39 | |
Dataset2 | Xgboost | 55.54 | 97.20 | 44.76 | 45.47 | 83.57 |
KNN | 57.16 | 98.43 | 68.46 | 58.93 | 91.01 | |
LightGBM | 47.11 | 96.99 | 52.57 | 53.24 | 88.21 | |
SVM | 55.27 | 98.38 | 58.91 | 58.78 | 91.12 | |
RF | 54.69 | 98.26 | 56.40 | 56.12 | 89.26 | |
Dataset3 | Xgboost | 62.19 | 96.79 | 52.13 | 51.82 | 84.78 |
KNN | 70.12 | 98.48 | 71.37 | 70.52 | 91.46 | |
LightGBM | 52.61 | 97.28 | 54.68 | 55.26 | 89.12 | |
SVM | 63.99 | 98.28 | 65.11 | 64.36 | 91.89 | |
RF | 68.63 | 98.44 | 66.98 | 66.77 | 90.68 |
