ELKI

ELKI is an open-source Java framework, developed by a research team at LMU Munich, focused on unsupervised learning research — particularly clustering and outlier detection — with a modular architecture that separates data mining algorithms from data management for independent, fair evaluation. It uses data index structures like the R*-tree to optimize performance at scale, documents implementations tied to 220+ academic publications, and includes benchmarking and visualization tools for comparing algorithms rigorously. The current stable release (0.8.0) is available via Maven and GitHub. Completely free, licensed under AGPLv3, explicitly free for scientific usage. Good fit for a researcher or graduate student doing clustering or outlier-detection research who needs a framework built for rigorous, reproducible algorithm comparison rather than a general-purpose ML library.