Skip to main content

Table 5 MNAR:mid missing values

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

 

MAE

RMSE

Method

M-value

B-value

M-value

B-value

(a) Healthy datasets

    

mean

0.053 ±0.001

0.051 ±0.001

0.082 ±0.001

0.076 ±0.001

impute.knn

0.041 ±0.002

0.050 ±0.004

0.067 ±0.005

0.085 ±0.010

softImpute

0.051 ±0.001

0.050 ±0.006

0.078 ±0.003

0.080 ±0.012

imputePCA

0.045 ±0.001

0.043 ±0.001

0.072 ±0.001

0.067 ±0.001

SVDmiss

0.052 ±0.001

0.043 ±0.001

0.081 ±0.002

0.069 ±0.002

missForest

0.044 ±0.001

0.042 ±0.001

0.068 ±0.001

0.064 ±0.001

methyLImp

0.044 ±0.001

0.040 ±0.001

0.068 ±0.001

0.064 ±0.001

(b) Disease datasets

    

mean

0.076 ±0.001

0.072 ±0.001

0.109 ±0.001

0.101 ±0.001

impute.knn

0.060 ±0.002

0.073 ±0.006

0.091 ±0.005

0.116 ±0.010

softImpute

0.075 ±0.002

0.072 ±0.010

0.108 ±0.003

0.111 ±0.021

imputePCA

0.066 ±0.001

0.062 ±0.001

0.098 ±0.001

0.091 ±0.001

SVDmiss

0.075 ±0.001

0.064 ±0.001

0.112 ±0.002

0.100 ±0.002

missForest

0.066 ±0.001

0.062 ±0.001

0.098 ±0.001

0.090 ±0.001

methyLImp

0.065 ±0.001

0.057 ±0.001

0.095 ±0.002

0.088 ±0.001

  1. 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