Skip to contents

Auto-train a texture model from an iconic_data object (internal)

Source: R/model_sensitivity.R
dot-auto_train_gan.Rd

Converts the user's iconic_data to the load_real_input_data() format and trains a GAN (or multivariate-normal fallback) on the resulting tidy frame. The trained model supplies realistic covariate and outcome texture to run_single_iteration(), and carries feature-level residual correlation matrices for the Y, M, and W panels so the simulation can inject correlated noise.

Usage

.auto_train_gan(data, epochs = 100)

Arguments

data

An iconic_data object.

epochs

GAN training epochs. Default 100.

Value

An iconic_gan object.