Table 6 MNAR:high missing values

From: Methylation data imputation performances under different representations and missingness patterns

	MAE		RMSE
	M-value	B-value	M-value	B-value
(a) Healthy datasets
mean	0.026 ±0.001^∗	0.026 ±0.001	0.044 ±0.001^∗	0.044 ±0.001
impute.knn	0.054 ±0.013^∗	0.092 ±0.020	0.103 ±0.022^∗	0.152 ±0.023
softImpute	0.028 ±0.002^∗	0.033 ±0.010	0.049 ±0.005^∗	0.063 ±0.026
imputePCA	0.022 ±0.001^∗	0.022 ±0.001	0.039 ±0.001	0.038 ±0.001
SVDmiss	0.032 ±0.001	0.025 ±0.001^∗	0.056 ±0.004	0.043 ±0.002^∗
missForest	0.023 ±0.001^∗	0.023 ±0.001	0.039 ±0.001	0.038 ±0.001^∗
methyLImp	0.027 ±0.001	0.022 ±0.001^∗	0.044 ±0.001	0.039 ±0.002^∗
(b) Disease datasets
mean	0.041 ±0.001^∗	0.043 ±0.001	0.069 ±0.002^∗	0.069 ±0.002
impute.knn	0.085 ±0.017^∗	0.134 ±0.025	0.148 ±0.023^∗	0.203 ±0.024
softImpute	0.044 ±0.005^∗	0.053 ±0.018	0.075 ±0.009^∗	0.098 ±0.043
imputePCA	0.035 ±0.001^∗	0.036 ±0.001	0.061 ±0.002^∗	0.061 ±0.002
SVDmiss	0.050 ±0.001	0.041 ±0.001^∗	0.084 ±0.002	0.073 ±0.003^∗
missForest	0.036 ±0.001^∗	0.038 ±0.001	0.062 ±0.002^∗	0.062 ±0.001
methyLImp	0.037 ±0.001	0.032 ±0.001^∗	0.062 ±0.002	0.057 ±0.002^∗

Average Mean Average Error (MAE) and Root Mean Square Error (RMSE) imputation performance ± standard deviation. For each method, the ^∗ symbol indicates the measure (either M-value or β-value) for which the Wilcoxon signed-rank test p-value is <0.05. Best results per metric with respect to the Wilcoxon signed-rank test are highlighted in bold

Back to article page