A Survey of Computational Intelligence Techniques in Protein Function Prediction

Table 4

Summary of computational intelligence (CI) techniques in prediction of signal peptides.

Reference	CI techniques	Performance	Datasets
[88]	ANN	Accuracy: 97%	Amino acid sequences
[89]	ANN	Accuracy: 97%	Amino acid sequences
[90]	Bidirectional recurrent NN	Accuracy: 97%	Amino acid sequences
[91]	OET-NN	Accuracy: 73.4%	Pseudoamino acid composition
[92]	ANN	Accuracy: 93%	Amino acid sequences
[93]	SVM	Accuracy: 97%	Pseudoamino acid composition
[94]	SVM	Sensitivity: 90.97% and selectivity: 97.42%	Position specific amino acid composition
[95]	Bayesian reasoning network	Accuracy: 97.73% for secretory and nonsecretory and 90.90% for signal peptide cleavage site	Sequence derived features