Fig. 3From: Evaluation of different approaches for missing data imputation on features associated to genomic dataDistribution of mean RMSE for each method across missing data scenarios. A. Distribution of the mean RMSE in a 5-column MCAR imputation. B. Distribution of the mean RMSE in a 2-column MAR imputation. C. Distribution of the mean RMSE in a 2-column MNAR imputation. Distribution of mean RMSE when using a mean-based approach is shown in gray for comparison. RMSE corresponds to root mean squared error, MCAR corresponds to missing completely at random, MNAR corresponds to missing not at random and MAR corresponds to missing at random and Am corresponds to Amelia. Vertical dotted lines represent the median value of the distributionBack to article page