Skip to contents

Workflow

uccdf follows four steps:

  1. validate the input table
  2. infer a simple schema
  3. build a distance view and a latent view
  4. 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.