Aplicación de redes neuronales para la modelización y generación de series hidrológicas

This research aims to simulate average monthly flow information in watersheds: Rio Guayllabamba River Quijo, Rio Chico, Rio Pindo also daily data Catamayo basin, and Lemon, by using neural networks artificial, to obtain new funds, taking as input rainfall and watershed historical flows. The proposed model presents a multi-layered network architecture (feedforward) comprising three layers (input, hidden and output), the number of input neurons and hidden vary by model. The training function for modeling the watershed regulation Bayesian backpropagation. The training phase and validation of artificial neural network models is performed using different amount of input data, Depending on the basin.

• Hlavní autor: Carpio Peláez, Karina Patricia (author)
• Médium: bachelorThesis
• Jazyk: spa
• Vydáno: 2014
• Témata: Cuencas hidrográficas; Redes neuronales artificiales; Caudales de las cuencas; Ingeniería Civil – Disertaciones académicas
• On-line přístup: http://dspace.utpl.edu.ec/handle/123456789/10722