Desarrollo e implementación de un modelo híbrido de LDS y Algoritmos Genéticos para la obtención de resúmenes lingüísticos en datos de creep
Fecha
2012-06-28
Autores
González Muro, Amaury
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central "Marta Abreu" de la Villas
Resumen
La Sumarización Lingüística de Datos (LDS por sus siglas en inglés de Linguistic Data Summarization) es una técnica de reciente desarrollo cuya utilización se ha incrementado en los últimos tiempos en correspondencia con el desarrollo y la demanda de los Sistemas de Apoyo a la Toma de Decisiones guiados por datos. Esta técnica posee limitantes relacionadas con el proceso de búsqueda de proposiciones lingüísticas que realiza. En tal sentido se ha propuesto en la literatura el empleo de meta heurísticas de búsqueda no exhaustivas que permitan hacer más eficiente este proceso.
En el presente trabajo se propone el desarrollo e implementación de un modelo híbrido de Algoritmos Genéticos (AG) con LDS para la obtención de resúmenes lingüísticos sobre datos de creep, propiedad mecánica de los aceros utilizada en el diseño de nuevas aleaciones. Como parte del proceso de diseño se analizan todos los aspectos del AG y se definen las mejores variantes de tres operadores genéticos incluyendo dos nuevos que son propuestos para resolver problemas específicos en la problemática abordada. Como parte de la investigación se desarrolló la implementación de una herramienta automatizada que utiliza plataformas de desarrollo existentes relacionadas con la temática abordada. Como resultado se obtiene un modelo híbrido final que resulta escalable y permite obtener soluciones de buena calidad.
The Linguistic Data Summarization is a technique of development which has been increasingly implemented in recent times, in accordance with the development and demand of the Decision Support Systems (DSS) driven by data. The limitations of this technique consist on the searching process of linguistics sentences that it carries out. In this sense, the employment of a heuristic approach of non-exhaustive search that allows a more efficient process has been proposed in the reviewed literature. In this thesis, it is proposed the development and implementation of a hybrid model of Genetic Algorithm with LDS to obtain linguistic summaries about creep data, which is data related to the mechanic property of the steels used in the design of new metal alloys. As part of the design process, all the aspects of the Genetic Algorithm are analyzed, and the best variants of genetic operators are defined, including two new proposed operators to solve specific problems on the discussed topic. The implementation of an automatized tool that employs existing development platform related to the studied issue was also developed. As the result, a final hybrid model that is scalable and allows obtaining good quality solutions is obtained.
The Linguistic Data Summarization is a technique of development which has been increasingly implemented in recent times, in accordance with the development and demand of the Decision Support Systems (DSS) driven by data. The limitations of this technique consist on the searching process of linguistics sentences that it carries out. In this sense, the employment of a heuristic approach of non-exhaustive search that allows a more efficient process has been proposed in the reviewed literature. In this thesis, it is proposed the development and implementation of a hybrid model of Genetic Algorithm with LDS to obtain linguistic summaries about creep data, which is data related to the mechanic property of the steels used in the design of new metal alloys. As part of the design process, all the aspects of the Genetic Algorithm are analyzed, and the best variants of genetic operators are defined, including two new proposed operators to solve specific problems on the discussed topic. The implementation of an automatized tool that employs existing development platform related to the studied issue was also developed. As the result, a final hybrid model that is scalable and allows obtaining good quality solutions is obtained.
Descripción
Palabras clave
Sumarización Lingüística de Datos, Linguistic Data Summarization, Sistemas de Apoyo, Toma de Decisiones, Algoritmos Genéticos, Modelo Híbrido, Datos de Creep