Welcome to PyMinimax’s documentation!

Package Status

Documentation Status Continuous Integration Coverage Status PyPI Downloads License

PyMinimax is a Python implementation of Bien and Tibshirani’s paper “Hierarchical Clustering With Prototypes via Minimax Linkage1. This is an agglomerative hierarchical clustering algorithm available in the protoclust R package 2 but not currently in SciPy nor scikit-learn. It has a great advantage over classical hierarchical clustering methods that in each cluster one prototype is selected from the original data. Thus the results are better interpretable.

PyMinimax has a SciPy-compatible API. Users who are familiar with the scipy.cluster.hierarchy module will find it easy to use.



J. Bien and R. Tibshirani. Hierarchical clustering with prototypes via minimax linkage. Journal of the American Statistical Association, 106(495), 2011, 1075-1084.


J. Bien and R. Tibshirani. Package ‘protoclust’, 2015.