Tabular and latent space synthetic data generation: a literature review

Fonseca, Joao; Bacao, Fernando

doi:10.1186/s40537-023-00792-7

Journal of Big Data

Table 2 Summary of the synthetic data generation methods discussed in this work. A field containing “—” indicates that the it is either not applicable to the corresponding method, and/or applies its own unique approach

From: Tabular and latent space synthetic data generation: a literature review

Algorithm	ML problem	Type	Architecture	Level	Data space	Scope
SDV [6]	Anon.	PDF	Probabilistic	External	Input	Global
MST [46]	DP	PGM	Probabilistic	External	Input	Global
MWEM [47]	DP	Other	Probabilistic	External	Input	Global
MWEM-PGM [48]	DP	PGM	Probabilistic	External	Input	Global
PrivBayes [49]	DP	PGM	Probabilistic	External	Input	Global
DPGAN [50]	DP	GAN	Network	External	Latent	Global
DPCTGAN [51]	DP	GAN	Network	External	Latent	Global
PATE-GAN [52]	DP	GAN	Network	External	Lat. + Out.	Global
PATECTGAN [51]	DP	GAN	Network	External	Lat. + Out.	Global
FEM [53]	DP	Perturb.	Probabilistic	External	Input	Global
RAP [54]	DP	Perturb.	Probabilistic	External	Input	Global
PDF [55, 56]	—	PDF	Probabilistic	External	Input	Global
Kamino [57]	DP	PDF	Probabilistic	External	Input	Global
RON-GAUSS [58]	DP	PDF	Probabilistic	Internal	Latent	Global
HDMM [59]	DP	Perturb.	Probabilistic	External	Input	Global
DualQuery [60]	DP	Other	Probabilistic	External	Input	Global
ROS(E) [61]	Ovs	Perturb.	Randomized	External	Input	Local
SMOTE [62]	Ovs	Linear	Randomized	External	Input	Local
SMOTENC [62]	Ovs	Linear	Randomized	External	Input	Local
SMOTEN [62]	Ovs	—	—	External	Input	Local
Borderline-SMOTE [63]	Ovs	Linear	Randomized	External	Input	Local
G-SMOTE [64]	Ovs	Geometric	Randomized	External	Input	Local
ADASYN [65]	Ovs	Linear	Randomized	External	Input	Local
KernelADASYN [66]	Ovs	PDF	Probabilistic	External	Input	Local
MOKAS [67]	Ovs	Other	Network	External	Latent	Global
SOMO [68]	Ovs	Linear	Net.+Rand.	External	Input	Global
G-SOMO [69]	Ovs	Geometric	Net.+Rand.	External	Input	Global
GMM-SENN [70]	Ovs	PDF	Probabilistic	External	Input	Global
GMF-SMOTE [71]	Ovs	Linear	Randomized	External	Input	Global
C-VAE [72]	Ovs	AE	Network	External	Latent	Global
Safe-level SMOTE [73]	Ovs	Linear	Randomized	External	Input	Local
LR-SMOTE [74]	Ovs	Linear	Randomized	External	Input	Global
K-means SMOTE [75]	Ovs	Linear	Randomized	External	Input	Global
DBSMOTE [76]	Ovs	Linear	Randomized	External	Input	Local
CGAN [77]	Ovs	GAN	Network	External	Latent	Global
K-means CTGAN [78]	Ovs	GAN	Network	External	Latent	Global
SMOTER [79]	Ovs + Reg	Linear	Randomized	External	Input	Local
G-SMOTER [80]	Ovs + Reg	Linear	Randomized	External	Input	Local
RACOG [81]	Ovs	PGM	Probabilistic	External	Input	Global
wRACOG [81]	Ovs	PGM	Probabilistic	External	Input	Global
RWO [82]	Ovs	PGM	Probabilistic	External	Input	Global
PDFOS [83]	Ovs	PDF	Probabilistic	External	Input	Global
Mixup [84]	DA	Linear	Randomized	External	In.+Out.	Local
M-Mixup [85]	DA	Linear	Network	Internal	Lat.+Out.	Global
NL-Mixup [86]	DA	Geometric	Randomized	External	In.+Out.	Local
AE-DA [87]	DA	AE	Network	External	In./Lat.+Out.	Local
MODALS [88]	DA	—	Network	Internal	Latent	Global
LSI [89]	DA	AE	Network	External	Lat.+Out.	Global
Gibbs [16]	DA	PGM	Probabilistic	External	Input	Global
MedGAN [90]	DA	GAN	Network	External	Latent	Global
GANBLR [91]	DA	PGM	Probabilistic	External	Input	Global
Table-GAN [92]	DA	GAN	Network	External	Latent	Global
CTGAN [93]	DA	GAN	Network	External	Latent	Global
TVAE [93]	DA	AE	Network	External	Latent	Global
AE [94]	DA	AE	Network	External	Latent	Global
InfoMixup [9]	AL	Linear	Network	Internal	Lat.+Out.	Global
VAEACGAN [95]	AL	AE	Network	Internal	Latent	Global
AL-G-SMOTE [44]	AL	Geometric	Randomized	Internal	Input	Local
DAE [96]	Semi-SL	AE	Network	Internal	Input	Global
\(\Pi\)-model [97]	Semi-SL	Perturb.	Randomized	Internal	In.+Lat.	Local
Mean Teacher [98]	Semi-SL	Perturb.	Randomized	Internal	In.+Lat.	Local
ICT [99]	Semi-SL	Linear	Randomized	Internal	Input	Local
Mixmatch [100]	Semi-SL	Linear	Randomized	Internal	Input	Local
SDAT [101]	Semi-SL	AE+PDF	Net.+Prob.	Internal	Latent	Global
MCoM [102]	Semi-SL	Linear	Randomized	Int.+Ext.	Inp.+Lat.	Global
C-Mixup [103]	Semi/Self-SL	AE+Lin.	Net+Rand.	Internal	Latent	Global
VIME [1]	Semi/Self-SL	Perturb.	Randomized	Internal	Input	Local
SubTab [104]	Self-SL	Perturb.	Rand.+Prob.	Internal	Input	Local
Scarf [105]	Self-SL	Perturb.	Randomized	Internal	Input	Local
A-SFS [106]	Self-SL	Perturb.	Randomized	Internal	Input	Local

Back to article page