Synthetic Tabular Data Generation on 90 MCAR Scenarios (6 datasets x 5 missing ratios) benchmark leaderboard · SOTA2 Research

Synthetic Tabular Data Generation on 90 MCAR Scenarios (6 datasets x 5 missing ratios)

18.1Alpha-PrecisionTVAE

Evaluation Results

Method	Links
TVAE2026.06		18.1	18.1	17.7	18.7	15.7	17.7
ARF2026.06		17.1	18.4	20.4	20	16.7	18.5
TabDiff2026.06		16.4	11	12.1	13.1	11.1	12.7
MissDiff2026.06		16.1	17.3	13.3	15.7	16.8	15.8
TabDDPM2026.06		15.5	12	15	13.5	11.4	13.5
TabSyn2026.06		13.6	13.7	13.3	14.9	13.2	13.7
CDTD2026.06		13.1	14.4	13.2	15.5	12.8	13.8
DiffPuter2026.06		12.7	13.2	12.8	12.8	17.2	13.7
CTGAN2026.06		12.4	16.7	15.9	16.5	17.6	15.8
CTGAN + AugFull2026.06		12.1	13.6	12.4	11.1	15	12.8
TVAE + AugFull2026.06		11.7	11.8	10.6	11	10.9	11.2
ForestDiff2026.06		10.7	5.4	8.4	9.5	6.4	8.1
ARF + AugFull2026.06		8.4	12.1	18.4	12.1	11	12.4
TabDDPM + AugMask2026.06		8	5.3	8.6	7.6	5.7	7
TabDDPM + AugFull2026.06		7.1	7.5	8.8	7.5	6.5	7.5
MIWAE2026.06		6.9	10	8.9	8.4	10.4	8.9
TabSyn + AugFull2026.06		6.8	7.6	5.3	5	6.5	6.3
TabDiff + AugFull2026.06		6.1	6.1	3.8	4	5.9	5.2
CDTD + AugFull2026.06		5.6	6.4	4.9	5.3	6.8	5.8
TabDiff + AugMask2026.06		4.7	3.3	2.3	3	3	3.2
CDTD + AugMask2026.06		4	4.2	3	3.5	4.4	3.8