Table 6 The DCR and NNDR on Ship-D and California House datasets.
From: A tabular data generation framework guided by downstream tasks optimization
Dataset | Strategy | DCR | NNDR | ||||
|---|---|---|---|---|---|---|---|
X+X’ | X | X’ | X+X’ | X | X’ | ||
Ship-D | RTVAE | 4.6912 | 3.8366 | 4.9217 | 0.8401 | 0.8589 | 0.8680 |
CTGAN | 4.7435 | 3.8366 | 4.7464 | 0.8434 | 0.8589 | 0.8049 | |
TabDDPM | 5.2821 | 3.8366 | 6.1570 | 0.7905 | 0.8589 | 0.9127 | |
DDPM with classifier | 5.3773 | 3.8366 | 6.0015 | 0.8359 | 0.8589 | 0.9066 | |
ShipGen | 5.7998 | 3.8366 | 5.9438 | 0.7985 | 0.8589 | 0.7443 | |
TDGGD | 5.3116 | 3.8366 | 4.9025 | 0.8326 | 0.8589 | 0.7959 | |
California | RTVAE | 0.5973 | 0.2549 | 0.2442 | 0.6304 | 0.6041 | 0.4051 |
House | CTGAN | 0.5181 | 0.2549 | 0.4870 | 0.6013 | 0.6041 | 0.5715 |
TabDDPM | 0.8066 | 0.2549 | 0.8986 | 0.4961 | 0.6041 | 0.5116 | |
DDPM with classifier | 0.8815 | 0.2549 | 1.6245 | 0.4924 | 0.6041 | 0.6026 | |
ShipGen | 1.1297 | 0.2549 | 1.3020 | 0.6071 | 0.6041 | 0.6393 | |
TDGGD | 1.3312 | 0.2549 | 0.0085 | 0.9406 | 0.6041 | 0.2777 | |
