Table 1.

Area under the learning curve (ALC) and number of training examples required to reach target area under the ROC curve (AUC) of the uncertainty, representative, and combined query strategies evaluated on the substance interactions and clinical medicine datasets

Type	Query strategy	Substance interactions		Clinical medicine
		ALC	|L| @ 0.80 AUC	ALC	|L| @ 0.80 AUC
Baseline	Passive	0.590	1295	0.491	2473
	SM	0.597	1218	0.541	2093
Uncertainty	LC	0.606	1051	0.543	2043
	LCB2	0.607	1060	0.542	2089
	D2C	0.623	891	0.548	2166
Representative	Density	0.622	905	0.547	2136
	Min-Max	0.634	657	0.550	2127
Combined	ID (β = 0.01)	0.626	771	0.534	2157
	ID (β = 1)	0.642	546	0.542	2146
	ID (β = 100)	0.635	653	0.550	2174
	ID (dynamic β)a	0.641	587	0.549	2180

Bold values indicate the best performing method for that metric.

^aNovel algorithm.