Fig. 2.

Performance of the active learning strategies in the HIV-1 protease inhibitors set. a Mean squared error at each iteration for the pool of regressors strategy vs. random selection and the full model. b Coefficient of determination at each iteration for the pool of regressors strategy vs. random selection and the full model. c Mean squeared error at each iteration for the distance to the training set strategy vs. random selection and the full model. d Coefficient of determination at each iteration for the distance to the training set strategy vs. random selection and the full model