Tabular Data Generation on Beijing (test)
0.423MLEReal
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| RealSource=Original Data2024.11 | 0.423 | — | — | — | |
| TabSynArchitecture=Diffusion-based2024.11 | 0.582 | 1.12 | — | — | |
| SMOTE2024.11 | 0.593 | 1.85 | — | — | |
| GReaTArchitecture=LLM-based2024.11 | 0.653 | 8.25 | — | — | |
| DiffLMArchitecture=Diffusion-based, Decoder=Mistral-v0.32024.11 | 0.696 | 6.35 | — | — | |
| TVAEArchitecture=VAE-based2024.11 | 0.77 | 19.16 | — | — | |
| CoDiArchitecture=Diffusion-based2024.11 | 0.818 | 16.94 | — | — | |
| Mistral-7B ICLProtocol=5-shot ICL2024.11 | 0.865 | 12.45 | — | — | |
| CTGANArchitecture=GAN-based2024.11 | 0.902 | 21.39 | — | — | |
| GPT-4 ICLProtocol=5-shot ICL2024.11 | 0.992 | 26.1 | — | — | |
| GOGGLEArchitecture=VAE-based2024.11 | 1.09 | 16.93 | — | — | |
| Qwen2.5-14B ICLProtocol=5-shot ICL2024.11 | 1.289 | 27.46 | — | — | |
| CTGAN2024.12 | — | — | 1.01 | — | |
| CTGANModel Category=GAN-based2026.04 | — | — | — | 96.27 | |
| GReaTModel Category=LLM-based2026.04 | — | — | — | 98.32 | |
| Real2024.12 | — | — | 0.423 | — | |
| TabCD2024.12 | — | — | 1.06 | — | |
| TabDDPM2024.12 | — | — | 0.556 | — | |
| TabDDPMModel Category=Diffusion-based2026.04 | — | — | — | 97.93 | |
| TabDiffModel Category=Diffusion-based2026.04 | — | — | — | 98.06 | |
| TabED-Cycperturbation=Cyc2024.12 | — | — | 1.01 | — | |
| TabED-Gridperturbation=Grid2024.12 | — | — | 0.978 | — | |
| TabED-Ordperturbation=Ord2024.12 | — | — | 1.02 | — | |
| TabED-Strperturbation=Str, feature knowledge=prior knowledge about structure2024.12 | — | — | 0.978 | — | |
| TabED-Uniperturbation=Uni2024.12 | — | — | 1.04 | — | |
| TabGRAAModel Category=LLM-based2026.04 | — | — | — | 99.51 | |
| TabSynModel Category=Diffusion-based2026.04 | — | — | — | 87.51 | |
| TVAE2024.12 | — | — | 1.05 | — | |
| TVAEModel Category=GAN-based2026.04 | — | — | — | 97.2 |