Carrot2 - Open Source Search Results Clustering Engine

Below is a summary of clustering algorithms that work within the Carrot² framework.


Algorithm	Author	Speed [s]*			Hierarchical clustering	Other features	Papers	Example results
Algorithm	Author	100	200	400	Hierarchical clustering	Other features	Papers	Example results
Lingo***	Stanislaw Osinski	0.48 0.16**	0.34 0.17**	0.74 0.31**	no		[2][3]	london
STC	Oren Zamir (impl: Dawid Weiss)	0.01	0.02	0.06	no		[5]	london
Lingo3G****	Stanisław Osiński (Carrot Search)	0.01	0.03	0.05	yes	multilingual clustering, synonyms, advanced tuning, scalability (5000 snippets in 1.3s*)		london

*) Clustering speed measurements were done for 100, 200, 400 snippets downloaded from Yahoo! for query 'lucene'. Benchmark environment: Intel Core2 Duo E8400 3GHz, 3GB MB RAM, Windows XP. Java Virtual Machine: Sun JDK 1.6.0, JVM switches: -server. Time presented in the table is an average of 100 runs, for each algorithm time measurement was preceded by 100 untimed warm-up runs.

**) Clustering time when native matrix computations are enabled.

***) Lingo is the default clustering algorithm used in the Carrot² live demos.

****) Lingo3G is a commercial document clustering engine and is not available in the Open Source part of Carrot². Please contact Carrot Search for details.


© 2002-2009 Stanislaw Osinski, Dawid Weiss, updated: 15 June 2009

Algorithms