Exploring new policies of metric access methods construction for the optimization of cluster detection algorithms

• Grant number: 11/16067-1
• Support Opportunities: Scholarships in Brazil - Master
• Start date: March 01, 2012
• End date: January 31, 2013
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Maria Camila Nardini Barioni
• Grantee: Jessica Andressa de Souza
• Host Institution: Centro de Matemática, Computação e Cognição (CMCC). Universidade Federal do ABC (UFABC). Ministério da Educação (Brasil). Santo André , SP, Brazil

Abstract

The significant growth in both the amount and variety of data available for analysis has increased the need to develop methodologies for understanding, processing and summarizing data in a quick and automatic manner. One of the features that has proven critical for understanding and learning from large amounts of data is related to the data organization into groups by synthesis of what is common. Clustering detection is among the techniques used in the data mining field to aid data analysis by synthesis. This technique aims to group the data into groups of similar elements based on measured or perceived data features.The development of such a technique has been extensively explored in the literature. However, there are at least two open issues: the algorithms are not scalable for large datasets and, moreover, as these algorithms strategy is based on an unsupervised process, the results do not always correspond to the most appropriate data partitioning. Thus, the proposed work presented in this project aims to explore these problems by addressing the study and development of strategies that support the implementation of scalable clustering detection algorithms that results in clusters with better semantic quality.For this purpose, we aim to explore the development of new policies of metric access methods construction in order to support clustering detection processes.