Automatic clustering with swarm intelligence algorithms

• Grant number: 16/24583-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: January 01, 2017
• End date: June 30, 2017
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Adriane Beatriz de Souza Serapião
• Grantee: João Gabriel Zupi Cattani
• Host Institution: Instituto de Geociências e Ciências Exatas (IGCE). Universidade Estadual Paulista (UNESP). Campus de Rio Claro. Rio Claro , SP, Brazil

Abstract

Clustering is a data mining problem that consists on providing a set of data, grouping its objects so that the objects that are more similar to each other can be in the same cluster and the least similar objects can be allocated in other clusters. In general, the number of clusters is previously defined for the separation of objects. However, in many real-world situations, the number of clusters is not known and can not even be estimated at the start. This problem is called automatic clustering. In this project, three Collective Intelligence algorithms will be used for the problem of automatic clustering of numerical data sets. Such methods will be used to maximize cluster compactness indices and to minimize intracluster distances in order to find the optimal cluster number and centroid positions. The bio-inspired optimization methods Particle Swarm Optimization, Fish School Search and Gray Wolf Optimization will be tailored to effect grouping using the partitioning approach. The results of the clustering algorithms with Collective Intelligence will be compared to each other and evaluated by internal and external validation indexes. (AU)