Table 11:

Gaussian and binary data with n = 200 and p = 1002 under MAR. Approximately 40% features and 88% cases contain missing values. Linear RR and MICE are not included due to running out of RAM. Detailed simulation setup information is in appendix.

Models	Style	Time(s)	Imp MSE	Imp accu	Bias(${\widehat{\beta }}_1$)	CR(${\widehat{\beta }}_1$)	SE(${\widehat{\beta }}_1$)	SD(${\widehat{\beta }}_1$)
SoftImpute	SI	21.8	0.0431	0.3331	−0.2793	0.64	0.2156	0.2263
GAIN	SI	98.9	0.8942	0.3331	0.8610	0.24	0.1588	0.7651
MIWAE	SI	57.7	0.1267	0.6785	0.6658	0.04	0.1617	0.2030
Sinkhorn	MI	41.6	0.1076	0.3278	0.4201	0.73	0.3036	0.2883
MI-NNGP1	MI	4.7	0.0116	0.6463	−0.0147	0.94	0.1492	0.1495
MI-NNGP1-BS	MI	3.4	0.0145	0.6188	−0.0115	0.98	0.1771	0.1436
MI-NNGP2	MI	3.9	0.0089	0.7289	0.0126	0.99	0.1470	0.1247
MI-NNGP2-BS	MI	9.5	0.0093	0.7006	−0.0014	0.98	0.1556	0.1258
Complete data	-	-	-	-	0.0156	0.96	0.1119	0.1041
Complete case	-	-	-	-	0.3856	0.70	0.2846	0.2937
ColMean Imp	SI	-	0.1255	0.3278	0.4643	0.71	0.3000	0.2362