Fangorn Forest (F2): a machine learning approach to classify genes and genera in the family Geminiviridae

BMC Bioinformatics

Table 4 Performance of the gene classification model using default parameters of Weka

Type of evaluation	ML algorithm	Weighted average among all classes
Type of evaluation	ML algorithm	Accuracy	Precision	Recall	F-Measure	MCC	AUC
Using a test set	MLP	0.972	0.973	0.973	0.972	0.968	0.985
	SMO	0.976	0.977	0.976	0.976	0.973	0.995
	RF	0.981	0.982	0.982	0.982	0.979	0.998
10-fold cross validation	MLP	0.970	0.971	0.971	0.971	0.967	0.994
	SMO	0.972	0.973	0.973	0.973	0.969	0.994
	RF	0.976	0.977	0.977	0.977	0.974	0.997
Leave-one-out	MLP	0.970	0.970	0.970	0.970	0.966	0.994
	SMO	0.9727	0.973	0.973	0.973	0.969	0.994
	RF	0.9759	0.976	0.976	0.976	0.973	0.997
Mean performance	MLP	0.9707	0.9713	0.9713	0.9710	0.9670	0.9910
	SMO	0.9736	0.9743	0.9740	0.9740	0.9703	0.9943
	RF	0.9776	0.9783	0.9783	0.9783	0.9753	0.9973

ISSN: 1471-2105