Table 1 Prediction of real-valued RSA using the GBRT algorithm based on five different sequence encoding schemes that incorporate various combinations of sequence features
From: PredRSA: a gradient boosted regression trees approach for predicting protein solvent accessibility
Feature | RMSE(%) | MAE(%) | PCC |
|---|---|---|---|
PSSM | 13.87 | 10.26 | 0.67 |
PSSM+DISO | 13.38 | 9.93 | 0.69 |
PSSM+DISO+SS | 13.04 | 9.64 | 0.71 |
PSSM+DISO+SS+SCE | 12.23 | 8.99 | 0.74 |
PSSM+DISO+SS+SCE+CS | 12.07 | 8.86 | 0.75 |