Research Article
ProFold: Protein Fold Classification with Additional Structural Features and a Novel Ensemble Classifier
Table 8
The accuracy of 5-fold cross validation on the features extracted from DD-dataset using 10 basic classifiers.
| Feature groups | Basic classifiers | Fivefold CV accuracy (%) |
| DSSP | LMT | 43.0 | RandomForest | 51.3 | LibSVM | 46.4 | SimpleLogistic | 43.0 | RotationForest | 49.7 | SMO | 36.4 | NaiveBayes | 43.4 | RandomTree | 32.8 | FT | 42.4 | SimpleCart | 37.7 |
| AAsCPC | LMT | 32.5 | RandomForest | 35.4 | LibSVM | 34.4 | SimpleLogistic | 32.5 | RotationForest | 27.7 | SMO | 34.4 | NaiveBayes | 28.3 | RandomTree | 11.6 | FT | 34.4 | SimpleCart | 20.6 |
| PSSM | LMT | 56.3 | RandomForest | 53.7 | LibSVM | 57.2 | SimpleLogistic | 55.9 | RotationForest | 56.1 | SMO | 30.2 | NaiveBayes | 42.4 | RandomTree | 29.6 | FT | 49.5 | SimpleCart | 33.4 |
| FunD | LMT | 42.1 | RandomForest | 43.1 | LibSVM | 21.2 | SimpleLogistic | 43.1 | RotationForest | 41.8 | SMO | 38.9 | NaiveBayes | 38.3 | RandomTree | 39.9 | FT | 44.1 | SimpleCart | 34.7 |
|
|
The basic classifier of each feature group with the highest accuracy.
|