The predictive performance of short-linear motif features in the prediction of calmodulin-binding proteins

BMC Bioinformatics

Table 3 Classification results for the score matrices with SLiMs obtained from CM using 3-fold cross validation

Dataset for classification	# features	Classifier	Accuracy (%)	MCC	ROC Area
S score matrix		SVM-Polynomial (c=1,g=0)	72.6	0.45	0.73
	100	RF	74.7	0.49	0.74
		k-NN (k=1)	78.3	0.57	0.78
T score matrix		SVM-Polynomial (c=1,g=0)	57.6	0.21	0.58
	100	RF	69.3	0.40	0.77
		k-NN (k=1)	58.1	0.26	0.58

ISSN: 1471-2105