Accurate and efficient target prediction using a potency-sensitive influence-relevance voter

Table 1 AUC performance in the cross-validation experiment on the ChEMBL dataset

Cutoff (\(\upmu \hbox {M}\))	MaxSim	MeanSim	11NN	IRV	PS-IRV	SVM	RF
All datasets
\(1\)	0.79	0.76	0.81	0.86	0.86	0.84	0.84
\(5\)	0.76	0.74	0.82	0.84	0.85	0.84	0.82
\(10\)	0.75	0.73	0.81	0.84	0.85	0.84	0.82
Datasets with fewer than 100 molecules
\(1\)	0.75	0.75	0.75	0.78	0.78	0.77	0.78
\(5\)	0.72	0.73	0.74	0.74	0.76	0.77	0.76
\(10\)	0.71	0.71	0.74	0.75	0.75	0.76	0.76
Datasets with more than 100 molecules
\(1\)	0.80	0.76	0.73	0.87	0.88	0.85	0.85
\(5\)	0.77	0.74	0.84	0.87	0.88	0.86	0.84
\(10\)	0.77	0.73	0.84	0.87	0.88	0.86	0.86
Datasets with more than 200 molecules
\(1\)	0.81	0.75	0.84	0.89	0.89	0.86	0.86
\(5\)	0.78	0.74	0.86	0.89	0.90	0.87	0.86
\(10\)	0.77	0.73	0.86	0.88	0.90	0.87	0.85

Each section of the table shows the average performance for datasets of different sizes
Best results within each group are in italics

ISSN: 1758-2946