1- Benchmark Suite for Clustering Algorithms - Version 1 (BETA)
2- ============================================================
1+ Benchmark Suite for Clustering Algorithms - Version 1
2+ =====================================================
33
44Maintained by [ Marek Gagolewski] ( http://www.gagolewski.com ) ,
55with contributions from [ Anna Cena] ( http://cena.rexamine.com )
66
77
88It is not rare for clustering papers/graduate theses to consider only few
9- (say, 5-10) datasets. Some authors propose own datasets, not
10- testing their methods against other benchmarks, risking their evaluation
11- be biased. Some authors who share their data (kudos to them!)
9+ (say, 5-10) datasets. Some authors just stick to 5-10 UCI datasets,
10+ which is not enough. Other authors propose own datasets, not
11+ testing their methods against other benchmarks batteries,
12+ risking their evaluation be biased. Some authors who share
13+ their data (kudos to them!)
1214might not necessarily make the use of their suites particularly
1315smooth (different file formats, different ways to access, etc.,
1416even across a single repository).
@@ -27,10 +29,9 @@ and data mining literature. Moreover, we introduce **new datasets**
2729of different dimensionalities, sizes and cluster types.
2830
2931
30- This is ** version 1** (BETA) of the suite,
31- dated XX YYYY 2020 (TODO).
32- The contents of this repository (TODO: will be) frozen --
33- no new datasets will be added here. This will ensure
32+ This is ** version 1** of the suite, dated 8 May 2020.
33+ The contents of this repository are frozen --
34+ no new datasets will be added. This will ensure
3435that everyone evaluates their methods on the same test data.
3536
3637
@@ -51,11 +52,13 @@ Please cite this repository as:
5152```
5253
5354
54- New releases/versions of this benchmark suite (as separate repositories)
55+ New releases/versions of this benchmark suite
56+ (as separate repositories)
5557are likely to appear in the future.
56- If you have any suggestions regarding its scope (e.g., request the addition
57- of new datasets),
58- please contact the maintainer, [ Marek Gagolewski] ( http://www.gagolewski.com ) .
58+ If you have any suggestions regarding its scope
59+ (e.g., request the addition of new datasets),
60+ please contact the maintainer,
61+ [ Marek Gagolewski] ( http://www.gagolewski.com ) .
5962
6063
6164The following is provided ** solely for research purposes** ,
0 commit comments