A tool for Multi-Strategy Learning

• Autor(a) principal: Francisco Reinaldo
• Data de Publicação: 2006
• Outros Autores: Marcus Siqueira, Rui Camacho, Luís Paulo Reis
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: https://hdl.handle.net/10216/67130
• topic: Tecnologia de computadores, Engenharia electrotécnica, electrónica e informática; Computer technology, Electrical engineering, Electronic engineering, Information engineering

This paper presents the AFRANCI tool for the development of Multi-Strategy learning systems. AFRANCI allows users to build, in an interactive and easy way, complex systems. Systems are built using a two step methodology: design of the structure of the system; and fill in the modules. The structure of the target system is a collection of interconnected modules. The user may then choose among a variety of learning algorithms to construct each module. The tool has several built-in Machine Learning algorithms and interfaces that enable it to use external learning tools like WEKA or CN2. AFRANCI uses the interdependency of the modules to determine the sequence of their training. To improve usability, the tool uses a wrapper that hides from the user the parameter tuning procedure for each algorithm. In a final step of the design sequence AFRANCI generates a compact and legible ready-to-use ANSI C++ open-source code for the final system. To illustrate the concept we have empirically evaluated the tool in the context of the RoboCup Rescue domain. We have developed a small system that uses both neural networks and rules in the same system. The experiment have shown that a very significant speed up is attained in the development of systems when using this tool.