A Framework for Benchmarking Clustering Algorithms

Results for Benchmark Suite Version 1

This repository is part of the Framework for Benchmarking Clustering Algorithms.

It hosts the outputs of various clustering algorithms for the datasets included in the Benchmark Suite for Clustering Algorithms: Version 1.1.0; see https://github.com/gagolews/clustering-data-v1/.

Author/Editor/Maintainer: Marek Gagolewski.

Refer to https://clustering-benchmarks.gagolewski.com for the full description, file format specification, example Python/R/MATLAB code, datasets explorer, and literature references.

The datasets are provided solely for research purposes. Contributions of new results are warmly encouraged.

Note that in the case of the slower, more memory-gluttonous methods, only the results for smaller datasets and a few cardinalities are provided.

Also, if a method fails to converge or returns a different number of clusters than requested, the respective output is not provided.

The directory named original gives the results on the original data spaces, i.e., where only some minor feature preprocessing was applied; see the README file therein for more details.

scale_standard contains some (incomplete) results for datasets with standardised variables.