Tyrosine Kinase Ligand-Receptor Pair Prediction by Using Support Vector Machine

Table 1

Prediction results of each method.

Method	Precision	Recall	-measure	AUC	TP	FP	TN	FN
1-mer frequency	—	—	—	0.638	0	0	2183	95
2-mer frequency	0.178	0.442	0.253	0.713	42	191	1992	53
Domain level	0.387	0.337	0.357	0.801	32	48	2135	63
Superfamily level	1.0	0.674	0.802	0.974	64	0	2183	31
Composite vector (2-mer frequency + superfamily level)	0.984	0.694	0.812	0.996	66	1	2182	29
Combined results (2-mer frequency + superfamily level)	0.868	0.611	0.712	0.906	58	10	2173	37

Each column in the table describes a single method. “Composite vector” describes the prediction based on the combination of feature vectors from 2-mer frequency method and domain, and “combined results” denotes the combination of the predicted results from 2-mer frequency method and domain (superfamily). TP, FP, TN, and FN: true positive, false positive, true negative, and false negative, respectively.