KinOrtho: a method for mapping human kinase orthologs across the tree of life and illuminating understudied kinases

Table 2 Performance of GO annotation prediction in each training set

GO domain	Model	Accuracy	Precision	Recall	F-measure	AUC
Biological process	Logistic regression	0.896	0.942	0.844	0.891	0.914
	SVM	0.896	0.943	0.843	0.890	0.896
	Random forest	0.909	0.944	0.869	0.905	0.923
Cellular component	Logistic regression	0.908	0.955	0.856	0.903	0.924
	SVM	0.908	0.955	0.858	0.903	0.908
	Random forest	0.921	0.959	0.880	0.918	0.932
Molecular function	Logistic regression	0.940	0.959	0.919	0.939	0.961
	SVM	0.942	0.959	0.925	0.941	0.942
	Random forest	0.955	0.964	0.945	0.955	0.965

The best performance in each measurement and each GO domain is highlighted in underlined

ISSN: 1471-2105