CopulaGANOptions<T>
Options & Configuration · AiDotNet.Models.Options
Configuration options for CopulaGAN, a synthetic tabular data generator that combines Gaussian copula transformations with the CTGAN training pipeline.
Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → FinancialNeuralNetworkOptions → RiskModelOptions<T> → SyntheticDataGeneratorOptions<T> → CopulaGANOptions<T>
For Beginners
CopulaGAN improves on CTGAN by “normalizing” each numerical column before training. Think of it like this:
- Copula Transform: Each number column is reshaped to look like a bell curve
(Gaussian distribution) using a mathematical trick called a “copula.”
- CTGAN Training: The standard CTGAN generator/discriminator pipeline trains
on this nicely-shaped data, which is easier to learn.
- Inverse Transform: Generated data is converted back to the original
distribution shapes.
This often produces better results than plain CTGAN for columns with unusual distributions (e.g., heavily skewed income data, bimodal distributions).
Example:
How It Works
CopulaGAN extends CTGAN by first transforming continuous columns using Gaussian copulas. This maps each continuous column’s distribution to a standard normal via CDF-then-quantile, making the CTGAN generator’s job easier and often improving generation quality for columns with complex or skewed distributions.
Reference: “Synthesizing Tabular Data using Copulas” (2020)
Constructors
| Constructor | Summary |
|---|---|
CopulaGANOptions() |
Initializes a new instance with the default settings. |
CopulaGANOptions(CopulaGANOptions<T> other) |
Initializes a new instance by copying every setting of another instance. |
Properties
| Property | Summary |
|---|---|
double DiscriminatorDropout { get; set; } |
Gets or sets the dropout rate for the discriminator. |
double GradientPenaltyWeight { get; set; } |
Gets or sets the gradient penalty weight (lambda) for WGAN-GP. |
double GumbelTemperature { get; set; } |
Gets or sets the Gumbel-softmax temperature of the generator’s mode and discrete-column heads. |
double LearningRate { get; set; } |
Gets or sets the learning rate for both generator and discriminator optimizers. |
int BatchSize { get; set; } |
Gets or sets the training batch size. |
int DiscriminatorSteps { get; set; } |
Gets or sets the number of discriminator training steps per generator step. |
int EmbeddingDimension { get; set; } |
Gets or sets the dimension of the random noise vector fed to the generator. |
int Epochs { get; set; } |
Gets or sets the number of training epochs. |
int PacSize { get; set; } |
Gets or sets the PacGAN packing size. |
int VGMModes { get; set; } |
Gets or sets the number of Gaussian mixture components for VGM normalization. |
int[] DiscriminatorDimensions { get; set; } |
Gets or sets the hidden layer sizes for the discriminator network. |
int[] GeneratorDimensions { get; set; } |
Gets or sets the hidden layer sizes for the generator network. |