Construcción de sistemas multiclasificadores usando Algoritmos Genéticos y medidas de diversidad
Fecha
2014-06-07
Autores
Morales Hernández, Alejandro
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
Las técnicas de clasificación están siendo muy utilizadas en la solución de diferentes problemas de la sociedad. Existen varios modelos de clasificación reportados en la literatura como las redes neuronales, árboles de clasificación y análisis discriminante. En investigaciones recientes muchos autores introducen el término multiclasificador como un “clasificador” que combina las salidas de un conjunto de clasificadores individuales, utilizando algún criterio (ej.; promedio, voto mayoritario, mínimo, etc.). Cuando se combinan clasificadores es importante garantizar la diversidad entre ellos ya que no tendría sentido combinar clasificadores cuya clasificación es la misma. Existen varios modelos para construir un multiclasificador y todos garantizan esta diversidad de diferentes formas. En el caso de aquellos que usan distintos clasificadores bases, existen algunas medidas estadísticas que pueden ser usadas para estimar cuán diversos son.
Además de los modelos tradicionales de construcción de multiclasificadores recientemente se están utilizando varias meta heurísticas con este propósito, dentro de las que se destacan los Algoritmos Genéticos. Están basados en el proceso genético que ocurre en los organismos vivos y en los principios de la evolución natural de las poblaciones. En este trabajo se presentan algunas medidas de diversidad y se implementa una variante de Algoritmo Genético con el objetivo de obtener una combinación de clasificadores diversos y una exactitud del sistema multiclasificador superior a la mejor exactitud individual. Además se realizan algunos experimentos para estudiar el comportamiento de la diversidad y las combinaciones de medidas que se proponen. Finalmente, se muestra una aplicación en el campo de la Bioinformática.
Classification techniques are being widely used in solving different problems in society. There exist several classification models referenced in the literature such as neural networks, classification trees and discriminant analysis. In recent research, many authors introduce the term “multiple classifier” as a "classifier" that combines the outputs of a set of individual classifiers using certain criteria (e.g., average, majority vote, minimum, etc.). When combining classifiers is important to ensure diversity among them, would not make sense to combine classifiers whose classification is the same. There are several models for constructing a multiple classifier system and all ensure the diversity of different ways. For those who use different base classifiers, there are some statistical measures that can be used to estimate how diverse they are. In addition to traditional construction multiple-classifier models, recently has begun using several meta heuristics, within which stand Genetic Algorithms. They are based on the genetic process occurring in living organisms and the principles of natural evolution of populations. In this work, some diversity measures are presented and variant genetic algorithm is implemented in order to obtain a combination of diverse classifiers and a multiple-classifier system's accuracy superior to the best individual accuracy. Also some experiments are performed to study the performance of diversity and combinations of measures proposed. Finally, an application is showed in the field of Bioinformatics.
Classification techniques are being widely used in solving different problems in society. There exist several classification models referenced in the literature such as neural networks, classification trees and discriminant analysis. In recent research, many authors introduce the term “multiple classifier” as a "classifier" that combines the outputs of a set of individual classifiers using certain criteria (e.g., average, majority vote, minimum, etc.). When combining classifiers is important to ensure diversity among them, would not make sense to combine classifiers whose classification is the same. There are several models for constructing a multiple classifier system and all ensure the diversity of different ways. For those who use different base classifiers, there are some statistical measures that can be used to estimate how diverse they are. In addition to traditional construction multiple-classifier models, recently has begun using several meta heuristics, within which stand Genetic Algorithms. They are based on the genetic process occurring in living organisms and the principles of natural evolution of populations. In this work, some diversity measures are presented and variant genetic algorithm is implemented in order to obtain a combination of diverse classifiers and a multiple-classifier system's accuracy superior to the best individual accuracy. Also some experiments are performed to study the performance of diversity and combinations of measures proposed. Finally, an application is showed in the field of Bioinformatics.
Descripción
Palabras clave
Algoritmos Genéticos, Medidas de Diversidad, Multiclasificado, Genetic Algorithms, Measures of Diversity, Multiple Classifiers