Variational Autoencoders for Heterogeneous Tabular Data
Builds the decoder graph for an AutoTab VAE
Extract decoder-only weights from a trained Keras model
Specifying Encoder and Decoder Architectures for VAE_train()
Rebuild the encoder graph to export z_mean and z_log_var
Extract encoder-only weights from a trained Keras model
Build the feat_dist data frame for AutoTab
Reorder feat_dist rows to match preprocessed data
Get the stored feature distribution
Get TensorFlow Addons module safely
Sample from the latent space
Min–max scale continuous variables
Mixture-of-Gaussians (MoG) prior in AutoTab
Reset all random seeds across R, TensorFlow, and Python
Set the feature distribution for AutoTab
Train an AutoTab VAE on mixed-type tabular data
Build and train a variational autoencoder (VAE) for mixed-type tabular data (continuous, binary, categorical). Models are implemented using 'TensorFlow' and 'Keras' via the 'reticulate' interface, enabling reproducible VAE training for heterogeneous tabular datasets.
Useful links