A new AntTree-based algorithm for clustering short-text corpora

• Main Authors: Marcelo Luis Errecalde, Diego Alejandro Ingaramo, Paolo Rosso
• Format: Article
• Language: English
• Published: Postgraduate Office, School of Computer Science, Universidad Nacional de La Plata 2010-04-01
• Series: Journal of Computer Science and Technology
• Subjects: internal validity measures; anttree; short-text clustering; bio-inspired algorithms; silhouette coefficient
• Online Access: https://journal.info.unlp.edu.ar/JCST/article/view/708

Research work on "short-text clustering" is a very important research area due to the current tendency for people to use "small-language", e.g. blogs, textmessaging and others. In some recent works, new bioinspired clustering algorithms have been proposed to deal with this difficult problem and novel uses of Internal Clustering Validity Measures have also been presented. In this work, a new AntTree-based approach is proposed for this task. It integrates information on the Silhouette Coefficient and the concept of attraction of a cluster in different stages of the clustering process. The proposal achieves results comparable to the best reported results in this area, showing an interesting stability in the quality of the results and presenting some interesting capabilities as a general improvement method for arbitrary clustering approaches.