Skip to main content
. 2022 Jul 26;94(31):10912–10920. doi: 10.1021/acs.analchem.1c04093

Figure 3.

Figure 3

Performance of imputation methods applied to various proportions of MAR, MNAR, and mixed (1:1) MAR:MNAR missingness simulated from the fungal and bacterial data sets, where smaller values indicate better relative performance. HM = half-minimum, BPCA = Bayesian principal components analysis, GSImp = Gibbs sampler imputation, QRILC = quantile regression imputation of left-censored data, and RF = Random Forest.