DataExplorer: a new approach to cluster analysis

• Grant number: 10/07059-2
• Support Opportunities: Research Grants - Young Investigators Grants
• Start date: April 01, 2011
• End date: March 31, 2015
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Katti Faceli
• Grantee: Katti Faceli
• Host Institution: Centro de Ciências e Tecnologias para a Sustentabilidade (CCTS). Universidade Federal de São Carlos (UFSCAR). Sorocaba , SP, Brazil
• City of the host institution: Sorocaba
• Associated researchers: André Carlos Ponce de Leon Ferreira de Carvalho ; Fabio Luciano Verdi ; Luciana Aparecida Martinez Zaina ; Marcilio Carlos Pereira de Souto ; Tiemi Christine Sakata

Abstract

Cluster analysis is commonly used in several areas. However, even the most recent approaches present dificulties that limit its use by experts of different areas. This research project is related to the development of a new approach to cluster analysis aiming at the obtaining of a greater variety of potentially useful clusters than that obtained by the traditional approaches or the more recent ensemble and multi-objective alternatives. The proposed approach aims at been a mechanism to obtain all (or most of) the relevant clusters hidden in one data set, regardless the structure they belong to (a partitition, for example). In order to achieve this, we look for clusters regardless the refinement level they occur, the clustering criteria they satisfy or if they are partially overlaped. In summary, the goal is to give the most complete possible description of the data in order to facilitate the work of the experts in the data, thus increasing the amount of knowledge they can extract using one single tool. (AU)