discoCVI-package: disco: Density-Informed Clustering Scoring and Optimisation

In discoCVI: Implementation of the DISCO Metric for Internal Clustering Evaluation

disco: Density-Informed Clustering Scoring and Optimisation

Description

The DISCO metric evaluates the quality of a clustering result using density-connectivity (DC) distances derived from the minimum spanning tree of the mutual-reachability graph. It generalises the silhouette coefficient to density-based partitions and assigns a principled score to noise points.

Author(s)

Maintainer: Amin Entezari [e-mail] (ORCID)

Authors:

• Anna Beer (Original Python implementation (arXiv:2503.00127))

• Lena Krieger (Original Python implementation (arXiv:2503.00127))

• Pascal Weber (Original Python implementation (arXiv:2503.00127))

Other contributors:

• Davide Chicco [e-mail] (ORCID) (Supervision and guidance) [contributor]

See Also

Useful links:

• https://github.com/aminentezari/discoCVI

• Report bugs at https://github.com/aminentezari/discoCVI/issues

discoCVI documentation built on Oct. 28, 2026, 1:07 a.m.