Evaluating a new feature selection strategy based on mutual information

• Grant number: 13/17320-8
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: November 01, 2013
• End date: January 31, 2014
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Maria Carolina Monard
• Grantee: Laís Pessine Do Carmo
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Feature Selection plays an important role in machine learning, as irrelevant and/or redundant features may reduce the quality and comprehensibility of the hypotheses induced by supervised learning algorithms. This work proposes the evaluation of a new feature selection strategy using mutual information, which will be estimated using a new method, recently proposed, related to the estimation of the ratio of two probability densities from a given collection of data. The algorithms developed using this new method to estimate mutual information will be experimentally evaluated, as well as compared with state-of-the-art feature selection algorithms.