Desarrollo e implementación de algoritmos paralelos aplicados a las extensiones de la teoría de los conjuntos aproximados (Rough Sets)
Archivos
Fecha
2008-07-20
Autores
Fernández Cancio, Jorge
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
El análisis de la información incompleta ha encontrado aplicación en muchas áreas, en particular la relacionada con la extracción de conocimiento de grandes conjuntos de datos. Las extensiones de la Teoría de los Conjuntos Aproximados se han aplicado con bastante éxito en el procesamiento de estos conjuntos de datos, aunque los mecanismos son muy diversos y no existe uno en particular que supere al resto, ofrecen actualmente una posibilidad de tratar este problema. El otro punto medular es el tema de la semejanza entre objetos En este trabajo se describen fundamentalmente dos algoritmos utilizados en el trabajo con información incompleta y uno empleado en el enfoque de la semejanza entre objetos usando las extensiones de la Teoría de los Conjuntos Aproximados, de estos se ofrecen variantes secuenciales y alternativas paralelos. Además se muestran los resultados obtenidos experimentalmente sobre una serie de Sistemas de Información almacenados en repositorios internacionales
The analysis of incomplete information has found applications in many areas particularly, in those related to the knowledge discovery of huge datasets. The extensions of the Theory of the Rough Sets have been successfully applied for processing this datasets and even though it exist many mechanisms, there is none, in particular, that we may say that outstrip the rest; they offer today a possibility to deal with this problem. Another important aspect is that of similarity among objects. In this work different algorithms are developed for the processing of incomplete information, and specifically of one based on the approach of the similarity among objects using the extensions of Theory of Rough Sets. Here we offer several serial variants and parallel alternatives that were experimentally tested on a set of information systems stored in international repositories.
The analysis of incomplete information has found applications in many areas particularly, in those related to the knowledge discovery of huge datasets. The extensions of the Theory of the Rough Sets have been successfully applied for processing this datasets and even though it exist many mechanisms, there is none, in particular, that we may say that outstrip the rest; they offer today a possibility to deal with this problem. Another important aspect is that of similarity among objects. In this work different algorithms are developed for the processing of incomplete information, and specifically of one based on the approach of the similarity among objects using the extensions of Theory of Rough Sets. Here we offer several serial variants and parallel alternatives that were experimentally tested on a set of information systems stored in international repositories.
Descripción
Palabras clave
Atributos Perdidos, Atributos Omitidos, Algoritmos Paralelos, Rough Sets, Sistema de Información Incompleto