CISS-VAE imputation result object

An S3 class returned by the CISS-VAE imputation function. Inherits from list, so all standard list operations (e.g., $, [[) work as expected. The class exists primarily to enable type-safe save/load methods that correctly handle the embedded PyTorch model.

Fields

The following fields may be present depending on which return_* arguments were set when the model was run:

imputed_dataset

Data frame. The full imputed dataset.

raw_data

Data frame. The original data before imputation.

model

Python object (CISSVAE). The trained PyTorch model. Present if return_model = TRUE.

cluster_dataset

Python object (ClusterDataset). Present if return_dataset = TRUE.

clusters

Integer vector. Cluster assignment per row. Present if return_clusters = TRUE.

silhouette_width

Numeric vector. Per-sample silhouette widths. Present if return_silhouettes = TRUE.

training_history

Data frame. Loss values recorded during training. Present if return_history = TRUE.

val_data

Data frame. Validation data with observed entries replaced by NA. Present if return_validation_dataset = TRUE.

val_imputed

Data frame. Model reconstructions of the validation entries. Present if return_validation_dataset = TRUE.

See also

save_impute_result, load_impute_result, print.impute_result, performance_by_cluster