Classifier combination rules applied to optimum-path forests

• Grant number: 13/12694-7
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2013
• End date: July 31, 2014
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Moacir Antonelli Ponti
• Grantee: Guilherme de Freitas Perinazzo
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Classifier combination methods are effective in many pattern recognition applications, under some conditions. The optimum-path forest classifier has been showing good potential for use as a base classifier. However, its original algorithm generates as output only the class labels, allowing fusion only by majority voting. In this project an adaptation is going to be made to generate as outputs decision profiles, that can be used to combine classifiers by rules such as mean, weighted mean, maximum, minimum, median and product. We expect to contribute by investigating this variant that have not been explored before. (AU)