Workflow
uccdf follows four steps:
- validate the input table
- infer a simple schema
- build a distance view and a latent view
- aggregate repeated clustering results across candidate
K
Design limits
The package is intentionally minimal.
- no large bundled data
- no external data downloads in package examples
- no advanced model backends
Intended use
The package is appropriate for lightweight exploratory clustering in
an R package setting where repository size matters.