Multi-Instance Multilabel Learning with Weak-Label for Predicting Protein Function in Electricigens

Table 5

Comparison results (mean ± std.) of MIMLwel models with four state-of-the-art MIML methods with different weak-label ratios on the Shewanella loihica PV-4 dataset.

W.L.R.	Methods	HL↓	maF1↑	miF1↑
20%	MIMLwel	0.013 ± 0.002	0.009 ± 0.008	0.145 ± 0.111
	MIMLNN	0.010 ± 0.002	0.000 ± 0.000	0.000 ± 0.000 ●
	MIMLRBF	0.011 ± 0.003	0.001 ± 0.001	0.001 ± 0.001 ●
	MIMLSVM	0.012 ± 0.002	0.005 ± 0.002	0.004 ± 0.002 ●
	EnMIMLNN {metric}	0.010 ± 0.003	0.001 ± 0.001	0.001 ± 0.001 ●
40%	MIMLwel	0.010 ± 0.002	0.005 ± 0.003	0.092 ± 0.039
	MIMLNN	0.010 ± 0.002	0.000 ± 0.000	0.000 ± 0.000 ●
	MIMLRBF	0.010 ± 0.002	0.001 ± 0.002	0.001 ± 0.002 ●
	MIMLSVM	0.012 ± 0.002	0.004 ± 0.002	0.004 ± 0.002 ●
	EnMIMLNN {metric}	0.010 ± 0.002	0.001 ± 0.003	0.001 ± 0.003 ●
60%	MIMLwel	0.011 ± 0.003	0.010 ± 0.006	0.167 ± 0.072
	MIMLNN	0.010 ± 0.003	0.001 ± 0.001	0.001 ± 0.001 ●
	MIMLRBF	0.010 ± 0.004	0.004 ± 0.004	0.003 ± 0.003 ●
	MIMLSVM	0.012 ± 0.003	0.005 ± 0.001	0.005 ± 0.002 ●
	EnMIMLNN {metric}	0.010 ± 0.003	0.005 ± 0.003	0.004 ± 0.003 ●
80%	MIMLwel	0.011 ± 0.003	0.011 ± 0.005	0.186 ± 0.043
	MIMLNN	0.010 ± 0.003	0.002 ± 0.001	0.001 ± 0.001 ●
	MIMLRBF	0.009 ± 0.003	0.008 ± 0.005	0.007 ± 0.005 ●
	MIMLSVM	0.012 ± 0.003	0.005 ± 0.002	0.005 ± 0.001 ●
	EnMIMLNN {metric}	0.010 ± 0.003	0.006 ± 0.004	0.005 ± 0.003 ●