Prediction of Protein-Protein Interactions from Protein Sequences by Combining MatPCA Feature Extraction Algorithms and Weighted Sparse Representation Models

Table 5

Comparison with the SVM-based method on three datasets.

Dataset	Classifier	ACC (%)	SN (%)	PE (%)	MCC (%)	AUC (%)
Yeast	WSRC	94.55 ± 0.63	97.15 ± 0.42	92.33 ± 0.91	89.68 ± 1.12	97.04 ± 0.24
	SVM	87.25 ± 0.66	88.22 ± 1.61	86.58 ± 1.71	77.75 ± 0.98	93.96 ± 0.41
Human	WSRC	97.48 ± 0.31	99.01 ± 0.32	96.27 ± 0.55	95.06 ± 0.60	99.05 ± 0.21
	SVM	92.88 ± 0.94	95.97 ± 0.74	90.91 ± 1.27	86.70 ± 1.63	97.56 ± 0.44
H. pylori	WSRC	83.64 ± 1.15	75.98 ± 1.63	89.71 ± 0.79	72.26 ± 1.49	86.35 ± 2.92
	SVM	82.64 ± 1.61	94.33 ± 1.87	76.51 ± 2.40	70.53 ± 2.38	91.60 ± 1.11