Biblioteca en Java para la aplicación de los conjuntos aproximados en el aprendizaje automatizado
Fecha
2013-07-04
Autores
Gómez Boix, Alejandro Geraldo
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
Se dice Aprendizaje Automatizado a aquellos métodos de solución de problemas de aprendizaje por las computadoras. Existen métodos de aprendizaje automatizado incorporados al Weka que utilizan la Teoría de los Conjuntos Aproximados (Rough Set Theory; RST) para su funcionamiento; sin embargo,
su incorporación al Weka incluye la implementación por separado de las facilidades de RST necesarias
para el funcionamiento de cada uno de ellos, ya que no existe una biblioteca que encapsule todos los
elementos de RST que se aplican en métodos de aprendizaje automatizado. El objetivo general del
trabajo de diploma consiste en desarrollar una biblioteca en Java que incorpore los principales
elementos de la RST clásica y extendida, aplicables en el desarrollo de métodos de aprendizaje automatizado y su validación. Los principales resultados obtenidos son: (1) la biblioteca que encapsula las definiciones, conceptos y medidas de RST aplicables en el desarrollo de métodos de Aprendizaje Automatizado, con la finalidad de incorporar esta biblioteca a la herramienta Weka, (2) la
implementación de un algoritmo para el cálculo de reductos que utiliza el modelo de optimización basado en colonias de hormigas y (3) la incorporación al Weka de un paquete para la validación del agrupamiento, el cual incorpora medidas de la Teoría de conjuntos aproximados, y por tanto, hace uso de la biblioteca creada.
It is called Machine Learning to those problem solving methods of learning by computers. There are machine learning methods incorporated to Weka that use the Rough Set Theory (RST) for its operation, however to join it to Weka includes the separated implementation of RST facilities that are necessary for the operation of each one of them, since there is no library that encapsulates all elements of RST applied in machine learning methods. The general objective of this research paper is to develop a Java library that incorporates the main elements of the classical and extended RST, applicable in the development of machine learning methods and its validation. The main results are: (1) the library that encapsulates the definitions, concepts and measures of RST, applicable in the development of Machine Learning methods, in order to incorporate this library to a Weka tool, (2) the implementation of a algorithm for calculating reducts that uses an optimization model based on ant colonies and (3) the incorporation to Weka of a package for clustering validation, which incorporates measures of rough sets theory, and thus makes use of the created library.
It is called Machine Learning to those problem solving methods of learning by computers. There are machine learning methods incorporated to Weka that use the Rough Set Theory (RST) for its operation, however to join it to Weka includes the separated implementation of RST facilities that are necessary for the operation of each one of them, since there is no library that encapsulates all elements of RST applied in machine learning methods. The general objective of this research paper is to develop a Java library that incorporates the main elements of the classical and extended RST, applicable in the development of machine learning methods and its validation. The main results are: (1) the library that encapsulates the definitions, concepts and measures of RST, applicable in the development of Machine Learning methods, in order to incorporate this library to a Weka tool, (2) the implementation of a algorithm for calculating reducts that uses an optimization model based on ant colonies and (3) the incorporation to Weka of a package for clustering validation, which incorporates measures of rough sets theory, and thus makes use of the created library.
Descripción
Palabras clave
Aprendizaje Automatizado, Biblioteca en Java, Modelo de Optimización, Colonias de Hormigas, Rough Set Theory (RST)