Improving the Multilayer Perceptron Learning by Using a Method to Calculate the Initial Weights with the Similarity Quality Measure Based on Fuzzy Sets and Particle Swarms
Título del documento: Improving the Multilayer Perceptron Learning by Using a Method to Calculate the Initial Weights with the Similarity Quality Measure Based on Fuzzy Sets and Particle Swarms
Revista: Computación y sistemas
Base de datos: PERIÓDICA
Número de sistema: 000383496
ISSN: 1405-5546
Autores: 1
1
1
2
Instituciones: 1Universidad de Camagüey, Departamento de Ciencias de la Computación, Camagüey. Cuba
2Universidad Central "Marta Abreu" de Las Villas, Departamento de Ciencias de la Computación, Santa Clara, Villa Clara. México
Año:
Periodo: Abr-Jun
Volumen: 19
Número: 2
Paginación: 309-320
País: México
Idioma: Inglés
Tipo de documento: Artículo
Enfoque: Analítico, descriptivo
Resumen en inglés The most widely used neural network model is Multilayer Perceptron (MLP), in which training of the connection weights is normally completed by a Back Propagation learning algorithm. Good initial values of weights bear a fast convergence and a better generalization capability even with simple gradient-based error minimization techniques. This work presents a method to calculate the initial weights in order to train the Multilayer Perceptron Model. The method named PSO+RST+FUZZY is based on the similarity quality measure proposed within the framework of the extended Rough Set Theory that employs fuzzy sets to characterize the domain of similarity thresholds. Sensitivity of BP to initial weights with PSO+RST+FUZZY was studied experimentally, showing better performance than other methods used to calculate feature weights
Disciplinas: Ciencias de la computación
Palabras clave: Redes,
Redes neuronales,
Perceptrón multicapa,
Conjuntos difusos,
Enjambre de partículas,
Medidas de similaridad
Keyword: Computer science,
Networks,
Neural networks,
Multilayer perceptron,
Fuzzy sets,
Particle swarm,
Similarity measures
Texto completo: Texto completo (Ver HTML)