Back to Search
Start Over
Improving GAN with inverse cumulative distribution function for tabular data synthesis
- Source :
- Neurocomputing. 456:373-383
- Publication Year :
- 2021
- Publisher :
- Elsevier BV, 2021.
-
Abstract
- Designing a generative model to synthesize realistic tabular data is of great significance in data science. Existing tabular data generative models have difficulty in handling complicated and diverse marginal distribution types due to the gradient vanishing problem, and these models pay little attention to the correlation between attributes. We propose a method that improves the generative adversarial network (GAN) with inverse cumulative distribution function for tabular data synthesis. This method first transforms continuous columns into uniform distribution data by using the cumulative distribution function, which can alleviate the gradient vanishing problem in model training. Then the method trains GAN with the transformed data, where the discriminator with label reconstruction function is presented to model the correlation among attributes accurately by introducing an auxiliary supervised task to help the correlations extraction. After that, we train a neural network for each continuous column to perform the inverse transformation of generated data into the target distribution, thereby the synthetic data is obtained. Experiments on simulated and real-world datasets show that our method compares favorably against the state-of-the-art methods in modeling tabular data.
- Subjects :
- 0209 industrial biotechnology
Uniform distribution (continuous)
Artificial neural network
Computer science
Cognitive Neuroscience
Cumulative distribution function
02 engineering and technology
Function (mathematics)
Synthetic data
Computer Science Applications
Generative model
020901 industrial engineering & automation
Transformation (function)
Artificial Intelligence
0202 electrical engineering, electronic engineering, information engineering
020201 artificial intelligence & image processing
Marginal distribution
Algorithm
Subjects
Details
- ISSN :
- 09252312
- Volume :
- 456
- Database :
- OpenAIRE
- Journal :
- Neurocomputing
- Accession number :
- edsair.doi...........775a24c1c2d8bfab2af8628b14581597