Skip to main content

Table 2 MCAR 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

0.030 ±0.001

0.030 ±0.001

0.051 ±0.001

0.050 ±0.001

impute.knn

0.039 ±0.007

0.059 ±0.012

0.079 ±0.015

0.112 ±0.019

softImpute

0.031 ±0.002

0.032 ±0.006

0.055 ±0.004

0.059 ±0.017

imputePCA

0.025 ±0.001

0.025 ±0.001

0.045 ±0.001

0.043 ±0.001

SVDmiss

0.035 ±0.001

0.027 ±0.001

0.063 ±0.002

0.048 ±0.002

missForest

0.026 ±0.001

0.026 ±0.001

0.044 ±0.003

0.043 ±0.002

methyLImp

0.029 ±0.001

0.025 ±0.001

0.050 ±0.002

0.047 ±0.002

(b) Disease datasets

    

mean

0.048 ±0.001

0.048 ±0.001

0.080 ±0.002

0.079 ±0.002

impute.knn

0.059 ±0.008

0.082 ±0.013

0.107 ±0.014

0.142 ±0.018

softImpute

0.050 ±0.004

0.051 ±0.010

0.084 ±0.007

0.091 ±0.026

imputePCA

0.041 ±0.001

0.041 ±0.001

0.072 ±0.002

0.070 ±0.002

SVDmiss

0.055 ±0.001

0.045 ±0.001

0.093 ±0.002

0.080 ±0.003

missForest

0.042 ±0.001

0.042 ±0.001

0.071 ±0.002

0.070 ±0.002

methyLImp

0.043 ±0.001

0.037 ±0.001

0.074 ±0.002

0.066 ±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