Nuevas propuestas en el ámbito de los operadores adaptativos para Sistemas Difusos Lingüísticos Evolutivos Multiobjetivo
Loading...
Publication date
Department
Research group
Center
Abstract
La presente investigación desarrolla nuevas propuestas de aprendizaje y ajuste en el ámbito de los operadores difusos adaptativos del Sistema de Inferencia y la Interfaz de Defuzzificación para Sistemas Difusos Lingüísticos Evolutivos Multiobjetivo. Para ello se presentan distintos algoritmos para el aprendizaje de Sistemas Basados en Reglas Difusas compactos y precisos que hace uso de ellos. Concretamente, se proponen modelos con buen equilibrio entre precisión e ¡nterpretabilidad, se avanza en el uso de medidas y diseño de índices que permitan cuantificar y optimizar mejor la ¡nterpretabilidad, y se proponen métodos para poder abordar el diseño de
sistemas difusos partiendo de conjuntos de datos de alta dimensionalidad y/o tamaño.
This research develops new learning and adjustment proposals in the field of operators used into adaptive fuzzy Inference System and defuzzification interface for Linguistic Multiobjective Evolutionary Fuzzy Systems. For this purpose different learning algorithms for obtain precise and compact Fuzzy Rule Based Systems have been designed and presented. Specifically, models with good trade-off between accuracy and interpretability are proposed, we progress in the use of measures and design indexes to quantify and better optimize interpretability, and methods to address the design of fuzzy systems based on data sets with high dimensional and / or size are proposed.
This research develops new learning and adjustment proposals in the field of operators used into adaptive fuzzy Inference System and defuzzification interface for Linguistic Multiobjective Evolutionary Fuzzy Systems. For this purpose different learning algorithms for obtain precise and compact Fuzzy Rule Based Systems have been designed and presented. Specifically, models with good trade-off between accuracy and interpretability are proposed, we progress in the use of measures and design indexes to quantify and better optimize interpretability, and methods to address the design of fuzzy systems based on data sets with high dimensional and / or size are proposed.
Keywords
Sistemas difusos; Modelado Difuso Lingüístico; Sistemas Difusos Evolutivos Multiobjetivo; Sistemas Basados en Reglas Difusas; Sistemas de Inferencia e Interfaz de Defuzzificación Adaptativa; Problemas de Regresión de Alta Dimensionalidad; Linguistic Fuzzy Modelling; Multi-Objective Evolutionary Fuzzy Systems; Fuzzy Rule Based Systems; Adaptive Inference Systems and Adaptive Defuzzification Interface; High-Dimensional Regression Problems














