Show HN: Dbgsom: A scikit-learn compatible Self-organizing Map

Category: library

Tags: self-organizing-map, clustering, manifold-learning

Score: 5.8/10 (Innovation: 5, Technical: 6, Documentation: 7, Utility: 5)

DBGSOM is a Python implementation of the Directed Batch Growing Self-Organizing Map, a neural network for clustering, classification, and data visualization that automatically determines the number of prototypes. It offers scikit-learn compatibility, built-in visualization, and competitive performance compared to classical SOMs and KMeans. The project is interesting for its dynamic topology-preserving map growth and integration into the sklearn ecosystem.

Target audience: data scientists, machine learning engineers

Repository: https://github.com/SandroMartens/DBGSOM · Jupyter Notebook · MIT · 4 stars

View on Hacker News