|
Ref. | CI techniques | Performance | Datasets |
|
[70] | -NN | Accuracy: 85% | Functional domain composition |
[71] | -NN | Accuracy: 76.6% | Amphiphilic pseudoamino acid composition |
[72] | OET-NN | Overall accuracy: 91.3%, 93.7%, and 98.3% for the 1st, 2nd, and 3rd level | Functional domain composition and PSSM |
[73] | -NN | Accuracy: 99% | Amino acid composition |
[74] | Fuzzy -NN | Accuracy: 56.9% | Pseudoamino acid composition, approximate entropy, and hydrophobicity |
[75] | SVM | Accuracy: 80.87% | Amphiphilic pseudo amino acid composition |
[76] | SVM with DWT | Accuracy: 91.9. | Pseudoamino acid composition |
[77] | SVM | MCC: 0.92 and accuracy: 93% | Pseudoamino acid composition with CTF |
[78] | SVM | Accuracy: 91.32% | Functional domain composition |
[79] | SVM | Accuracy: 81% to 98% and MCC: 0.82 to 0.98 | Pseudoamino acid composition with CTF |
[20] | SVM | Accuracy: 95.25% | Structural features based on fragment libraries |
[80] | SVM | Accuracy: 69.1–99.6% | Amino acid sequence |
[81] | SVM | Sensitivity: 85.6% and specificity: 86.1% | Pseudoamino acid composition |
[82] | SVM | Accuracy: 77.4% | Sequence similarity, amino acid composition, physiochemical properties, and dipeptide composition |
[83] | Bayesian classifier | Accuracy: 45% | Structural properties |
[84] | Random forest | Overall accuracy: 94.87%, 87.7%, and 84.25% for the 1st, 2nd, and 3rd level | Sequence derived features |
[85] | Random forest | Precision: 0.98 and recall: 0.89 | Set of specificity determining residues |
[86] | SVM and random forest | Accuracy: 71.29–99.53% by SVM and 94–99.31% by random forest | Sequence derived properties |
[87] | N-to-1 neural network | Overall accuracy: 96%, specificity: 80%, and FP rates: 7% | Amino acid sequences |
|