Skip to content

HDBSCAN

The hdbscan library is a suite of tools to use unsupervised learning to find clusters, or dense regions, of a dataset. The primary algorithm is HDBSCAN* as proposed by Campello, Moulavi, and Sander. The library provides a high performance implementation of this algorithm, along with tools for analysing the resulting clustering.

homepage: http://hdbscan.readthedocs.io/en/latest/

Available installations

HDBSCAN version Supported CPU targets Supported GPU targets EESSI version Module
0.8.38.post1 generic: aarch64, x86_64
Arm: a64fx, neoverse_n1, neoverse_v1, nvidia/grace
AMD: zen2, zen3, zen4
Intel: haswell, skylake_avx512, sapphirerapids, icelake, cascadelake
(none) 2023.06 HDBSCAN/0.8.38.post1-foss-2023a