Distribution based nearest neighbor imputation for truncated high dimensional data with applications to pre-clinical and clinical metabolomics studies

Table 4 Average MLCI of 100 simulations using the Myocardial, Atherothrombotic and African Race dataset with 15% missingness

Imputation Method	Myocardial Dataset	Atherothrombotic Dataset	African Race Dataset
Imputation Method	MLCI	MLCI	MLCI
Zero	0.061 (0.054)	0.086 (0.026)	0.028 (0.026)
Min	0.396 (0.061)	0.248 (0.069)	0.135 (0.078)
Mean	0.440 (0.089)	0.368 (0.103)	0.266 (0.134)
KNN-TR	0.510 (0.097)	0.392 (0.110)	0.274 (0.140)
KNN-CR	0.504 (0.094)	0.391 (0.110)	0.274 (0.139)
KNN-EU	0.474 (0.091)	0.380 (0.110)	0.272 (0.139)

The Myocardial dataset was comparing cases and controls of sample size 25 in each group, Atherothrombotic dataset was comparing sCAD and Type 2 MI of sample size 50 in each group and the African Race dataset was comparing Tanzania and Ethiopia of sample size 15 and 25 respectively

ISSN: 1471-2105