Skip to main content

Table 3 MAR 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

0.041 ±0.001

0.040 ±0.001

0.073 ±0.002

0.070 ±0.001

impute.knn

0.043 ±0.004

0.061 ±0.009

0.082 ±0.009

0.110 ±0.015

softImpute

0.042 ±0.002

0.043 ±0.007

0.077 ±0.005

0.082 ±0.017

imputePCA

0.037 ±0.001

0.036 ±0.001

0.069 ±0.002

0.066 ±0.002

SVDmiss

0.043 ±0.001

0.036 ±0.001

0.079 ±0.003

0.067 ±0.002

missForest

0.035 ±0.001

0.035 ±0.001

0.064 ±0.002

0.061 ±0.002

methyLImp

0.037 ±0.001

0.033 ±0.001

0.068 ±0.002

0.063 ±0.002

(b) Disease datasets

    

mean

0.060 ±0.001

0.060 ±0.001

0.101 ±0.002

0.097 ±0.002

impute.knn

0.067 ±0.005

0.087 ±0.010

0.115 ±0.009

0.144 ±0.014

softImpute

0.062 ±0.003

0.065 ±0.011

0.106 ±0.006

0.116 ±0.026

imputePCA

0.054 ±0.001

0.053 ±0.001

0.095 ±0.002

0.090 ±0.002

SVDmiss

0.067 ±0.001

0.057 ±0.001

0.114 ±0.003

0.104 ±0.004

missForest

0.053 ±0.001

0.053 ±0.001

0.093 ±0.002

0.088 ±0.002

methyLImp

0.053 ±0.001

0.049 ±0.001

0.092 ±0.002

0.089 ±0.002

  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