hepta_pdfFundamental Clustering Problem Suite

The Fundamental Clustering Problems Suite (FCPS) offers a variety of clustering problems any algorithm shall be able to handle when facing real world data. FCPS serves as an elementary benchmark for clustering algorithms.

FCPS consists of data sets with known a priori classifications that are to be reproduced by the algorithm. All data sets are intentionally created to be simple and might be visualized in two or three dimensions. Each data sets represents a certain problem that is solved by known clustering algorithms with varying success. This is done in order to reveal benefits and shortcomings of algorithms in question. Standard clustering methods, e.g. single-linkage, ward und k-means, are not able to solve all FCPS problems satisfactorily.

FCPS is supposed to be used in scientific works for free, as long as it is quotet as follows:

Ultsch, A.: Clustering with SOM: U*C, In Proc. Workshop on Self-Organizing Maps, Paris, France, (2005) , pp. 75-82

click here for data (.zip file)

Name	Problem	Description
Hepta	clearly defined clusters, different variances	Size 212 Dimensions 3 Classes 7
Lsun	different variances and inter cluster distances	Size 400 Dimensions 2 Classes 3
Tetra	almost touching clusters	Size 400 Dimensions 3 Classes 4
Chainlink	linear not separable	Size 1000 Dimensions 3 Classes 2
Atom	different variances and linear not separable	Size 800 Dimensions 3 Classes 2
EngyTime	Gaussian mixture	Size 4096 Dimensions 2 Classes 2
Target	outliers	Size 770 Dimensions 2 Classes 6
TwoDiamonds	cluster borders defined by density	Size 800 Dimensions 2 Classes 2
Wingnut	density vs. distance	Size 1070 Dimensions 2 Classes 2
Golfball	no clusters at all	Size 4002 Dimensions 3 Classes 2