Skip to main content
. 2015 Apr 30;1(2):206–215. doi: 10.1016/j.dadm.2015.01.006

Fig. 1.

Fig. 1

(A) Proportion of missing data for each feature, separately for mild cognitive impairment (MCI)-converters and MCI-stable patients. (B) Mean squared error between true and imputed values for different percentages of missing data. (C) Balanced accuracy for the different classification algorithms and different percentages of missing data.