Implementation of multiple-instance learning in drug activity prediction

BMC Bioinformatics

Table 7 Predictive performance for different models

Data set	Methods	Training set		Test set
		Accuracy	MCC	Accuracy	MCC
I	MILES^a	0.941	0.881	0.861	0.725
	Decision tree	0.915	0.830	0.781	0.569
	1-norm SVM	1.000	1.000	0.832	0.668
	Random forest	0.995	0.990	0.891	0.783
II	MILES^a	0.978	0.956	0.904	0.807
	Decision tree	0.955	0.913	0.919	0.837
	1-norm SVM	0.980	0.961	0.882	0.765
	Random forest	0.945	0.896	0.868	0.754
III	MILES^b	0.947	0.885	0.846	0.711
	Decision tree	0.966	0.924	0.838	0.682
	1-norm SVM	0.995	0.988	0.812	0.624
	Random forest	0.982	0.959	0.855	0.717
IV	MILES^b	0.898	0.811	0.794	0.584
	Decision tree	0.914	0.829	0.698	0.398
	1-norm SVM	0.952	0.906	0.714	0.418
	Random forest	0.936	0.877	0.698	0.392

ISSN: 1471-2105