Variational Autoencoders for Heterogeneous Tabular Data

Documentation for package ‘autotab’ version 0.1.1

decoder_model	Builds the decoder graph for an AutoTab VAE
Decoder_weights	Extract decoder-only weights from a trained Keras model
encoder_decoder_information	Specifying Encoder and Decoder Architectures for 'VAE_train()'
encoder_latent	Rebuild the encoder graph to export z_mean and z_log_var
Encoder_weights	Extract encoder-only weights from a trained Keras model
extracting_distribution	Build the 'feat_dist' data frame for AutoTab
feat_reorder	Reorder 'feat_dist' rows to match preprocessed data
get_feat_dist	Get the stored feature distribution
Latent_sample	Sample from the latent space
min_max_scale	Min–max scale continuous variables
mog_prior	Mixture-of-Gaussians (MoG) prior in AutoTab
reset_seeds	Reset all random seeds across R, TensorFlow, and Python
set_feat_dist	Set the feature distribution for AutoTab
VAE_train	Train an AutoTab VAE on mixed-type tabular data