Tabular and latent space synthetic data generation: a literature review

被引:33
作者
Fonseca, Joao [1 ]
Bacao, Fernando [1 ]
机构
[1] Univ Nova Lisboa, NOVA Informat Management Sch NOVA IMS, Campus Campolide, P-1070312 Lisbon, Portugal
关键词
Synthetic Data; Tabular data; Data privacy; Regularization; Oversampling; Active Learning; Semi-supervised Learning; Self-supervised Learning; UTILITY; SMOTE; ALGORITHM;
D O I
10.1186/s40537-023-00792-7
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
The generation of synthetic data can be used for anonymization, regularization, oversampling, semi-supervised learning, self-supervised learning, and several other tasks. Such broad potential motivated the development of new algorithms, specialized in data generation for specific data formats and Machine Learning (ML) tasks. However, one of the most common data formats used in industrial applications, tabular data, is generally overlooked; Literature analyses are scarce, state-of-the-art methods are spread across domains or ML tasks and there is little to no distinction among the main types of mechanism underlying synthetic data generation algorithms. In this paper, we analyze tabular and latent space synthetic data generation algorithms. Specifically, we propose a unified taxonomy as an extension and generalization of previous taxonomies, review 70 generation algorithms across six ML problems, distinguish the main generation mechanisms identified into six categories, describe each type of generation mechanism, discuss metrics to evaluate the quality of synthetic data and provide recommendations for future research. We expect this study to assist researchers and practitioners identify relevant gaps in the literature and design better and more informed practices with synthetic data.
引用
收藏
页数:37
相关论文
共 145 条
[1]  
ACKLEY DH, 1985, COGNITIVE SCI, V9, P147
[2]  
Alaa AM, 2022, PR MACH LEARN RES, P290
[3]   An Investigation of Credit Card Default Prediction in the Imbalanced Datasets [J].
Alam, Talha Mahboob ;
Shaukat, Kamran ;
Hameed, Ibrahim A. ;
Luo, Suhuai ;
Sarwar, Muhammad Umer ;
Shabbir, Shakir ;
Li, Jiaming ;
Khushi, Matloob .
IEEE ACCESS, 2020, 8 :201173-201198
[4]   A K-means Improved CTGAN Oversampling Method for Data Imbalance Problem [J].
An, Chunsheng ;
Sun, Jingtong ;
Wang, Yifeng ;
Wei, Qingjie .
2021 IEEE 21ST INTERNATIONAL CONFERENCE ON SOFTWARE QUALITY, RELIABILITY AND SECURITY (QRS 2021), 2021, :883-887
[5]  
Arik SO, 2021, AAAI CONF ARTIF INTE, V35, P6679
[6]   MedGAN: Medical image translation using GANs [J].
Armanious, Karim ;
Jiang, Chenming ;
Fischer, Marc ;
Kuestner, Thomas ;
Nikolaou, Konstantin ;
Gatidis, Sergios ;
Yang, Bin .
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS, 2020, 79
[7]  
Assefa SA, 2020, P 1 ACM INT C AI FIN, P1, DOI [10.1145/3383455.3422554, DOI 10.1145/3383455.3422554]
[8]  
Aydore S, 2021, PR MACH LEARN RES, V139
[9]  
Bahri D, 2022, INT C LEARN REPR 202
[10]   Deep learning: a statistical viewpoint [J].
Bartlett, Peter L. ;
Montanari, Andrea ;
Rakhlin, Alexander .
ACTA NUMERICA, 2021, 30 :87-201