Bibliotecas PUCV

Predicción de captura de anchovetas utilizando redes neuronales

Repositorio Dspace/Manakin

Buscar en DSpace


Búsqueda avanzada

Mostrar el registro sencillo del ítem

dc.contributor Rodríguez Agurto, José Nibaldo
dc.creator Pérez Galleguillos, Rodrigo Cristián
Fecha Ingreso dc.date.accessioned 2021-10-19T19:30:50Z
Fecha Disponible dc.date.available 2021-10-19T19:30:50Z
Fecha en Repositorio dc.date.issued 2021-10-19
Resumen dc.description <p>El&nbsp; problema&nbsp; de&nbsp; pron&oacute;stico&nbsp; de&nbsp; captura&nbsp; de&nbsp; anchovetas&nbsp; en&nbsp; la&nbsp; zona&nbsp; norte&nbsp; de&nbsp; Chile tradicionalmente ha sido modelado utilizando regresi&oacute;n lineal cl&aacute;sica (RLC). Sin embargo, la t&eacute;cnica&nbsp; de&nbsp; RLC&nbsp; no&nbsp; considera&nbsp; los&nbsp; fen&oacute;menos&nbsp; no&nbsp; lineales&nbsp; que&nbsp; puede&nbsp; exhibir&nbsp; el&nbsp; proceso&nbsp; de pron&oacute;stico.&nbsp; Por&nbsp; lo&nbsp; tanto,&nbsp; en&nbsp; esta&nbsp; memoria&nbsp; de&nbsp; t&iacute;tulo&nbsp; es&nbsp; propuesto&nbsp; un&nbsp; modelo&nbsp; no&nbsp; lineal&nbsp; de predicci&oacute;n&nbsp; de&nbsp; captura&nbsp; de&nbsp; anchovetas&nbsp; basado&nbsp; en&nbsp; redes&nbsp; neuronales&nbsp; sigmoidales&nbsp; (RNS).&nbsp; La arquitectura&nbsp; del&nbsp; modelo&nbsp; RNS&nbsp; est&aacute;&nbsp; compuesta&nbsp; por&nbsp; una&nbsp; capa&nbsp; de&nbsp; entrada,&nbsp; una&nbsp; capa&nbsp; oculta&nbsp; no lineal y una capa de salida lineal y los pesos de la RNS son estimados utilizando el algoritmo de&nbsp; aprendizaje&nbsp; Levenberg&nbsp; Marquardt.&nbsp; La&nbsp; mejor&nbsp; topolog&iacute;a&nbsp; encontrada&nbsp; durante&nbsp; la&nbsp; fase&nbsp; de evaluaci&oacute;n ha sido una RNS con seis nodos de entrada, diecis&eacute;is nodos ocultos y un nodo de salida. Adem&aacute;s, la varianza obtenida fue igual a un 91% con el modelo RNS propuesto.</p>
Resumen dc.description <p>The capture forecast problem of anchovy in the northern part of Chile traditionally has been shaped utilizing classical lineal decline (RLC). Nevertheless, the technique of RLC does not consider the not lineal phenomenon that can exhibit the process of forecast. Therefore, in this memory of title is proposed a not lineal model of prediction of capture of anchovy based on networks&nbsp; neural&nbsp; sigmoidal&nbsp; (RNS).&nbsp; The&nbsp; architecture&nbsp; of&nbsp; the&nbsp; model&nbsp; RNS&nbsp; is&nbsp; composed&nbsp; by&nbsp; an input layer, a not lineal hidden layer and the output lineal layer and the weights of the RNS are&nbsp; reckoned&nbsp; utilizing&nbsp; the&nbsp; algorithm&nbsp; of&nbsp; learning&nbsp; Levenberg&nbsp; Marquardt.&nbsp; The&nbsp; best&nbsp; topology found during the phase of evaluation has been a RNS with six nodes of input, sixteen hidden nodes and one node at the output. Besides, the variance obtained was equal to a 91% with the model RNS proposed.</p>
Resumen dc.description last modification
Resumen dc.description Licenciado en Ciencias de la Ingeniería
Resumen dc.description Ingeniero Civil en Informáticatítulo
Resumen dc.description INGENIERIA CIVIL INFORMATICA
Resumen dc.description <p>El&nbsp; problema&nbsp; de&nbsp; pron&oacute;stico&nbsp; de&nbsp; captura&nbsp; de&nbsp; anchovetas&nbsp; en&nbsp; la&nbsp; zona&nbsp; norte&nbsp; de&nbsp; Chile tradicionalmente ha sido modelado utilizando regresi&oacute;n lineal cl&aacute;sica (RLC). Sin embargo, la t&eacute;cnica&nbsp; de&nbsp; RLC&nbsp; no&nbsp; considera&nbsp; los&nbsp; fen&oacute;menos&nbsp; no&nbsp; lineales&nbsp; que&nbsp; puede&nbsp; exhibir&nbsp; el&nbsp; proceso&nbsp; de pron&oacute;stico.&nbsp; Por&nbsp; lo&nbsp; tanto,&nbsp; en&nbsp; esta&nbsp; memoria&nbsp; de&nbsp; t&iacute;tulo&nbsp; es&nbsp; propuesto&nbsp; un&nbsp; modelo&nbsp; no&nbsp; lineal&nbsp; de predicci&oacute;n&nbsp; de&nbsp; captura&nbsp; de&nbsp; anchovetas&nbsp; basado&nbsp; en&nbsp; redes&nbsp; neuronales&nbsp; sigmoidales&nbsp; (RNS).&nbsp; La arquitectura&nbsp; del&nbsp; modelo&nbsp; RNS&nbsp; est&aacute;&nbsp; compuesta&nbsp; por&nbsp; una&nbsp; capa&nbsp; de&nbsp; entrada,&nbsp; una&nbsp; capa&nbsp; oculta&nbsp; no lineal y una capa de salida lineal y los pesos de la RNS son estimados utilizando el algoritmo de&nbsp; aprendizaje&nbsp; Levenberg&nbsp; Marquardt.&nbsp; La&nbsp; mejor&nbsp; topolog&iacute;a&nbsp; encontrada&nbsp; durante&nbsp; la&nbsp; fase&nbsp; de evaluaci&oacute;n ha sido una RNS con seis nodos de entrada, diecis&eacute;is nodos ocultos y un nodo de salida. Adem&aacute;s, la varianza obtenida fue igual a un 91% con el modelo RNS propuesto.</p>
Resumen dc.description <p>The capture forecast problem of anchovy in the northern part of Chile traditionally has been shaped utilizing classical lineal decline (RLC). Nevertheless, the technique of RLC does not consider the not lineal phenomenon that can exhibit the process of forecast. Therefore, in this memory of title is proposed a not lineal model of prediction of capture of anchovy based on networks&nbsp; neural&nbsp; sigmoidal&nbsp; (RNS).&nbsp; The&nbsp; architecture&nbsp; of&nbsp; the&nbsp; model&nbsp; RNS&nbsp; is&nbsp; composed&nbsp; by&nbsp; an input layer, a not lineal hidden layer and the output lineal layer and the weights of the RNS are&nbsp; reckoned&nbsp; utilizing&nbsp; the&nbsp; algorithm&nbsp; of&nbsp; learning&nbsp; Levenberg&nbsp; Marquardt.&nbsp; The&nbsp; best&nbsp; topology found during the phase of evaluation has been a RNS with six nodes of input, sixteen hidden nodes and one node at the output. Besides, the variance obtained was equal to a 91% with the model RNS proposed.</p>
Formato dc.format PDF
Lenguaje dc.language spa
dc.rights sin documento
dc.source http://opac.pucv.cl/pucv_txt/txt-8000/UCI8380_01.pdf
Materia dc.subject Algoritmos
Materia dc.subject Regresión lineal
Materia dc.subject CAPTURA
Materia dc.subject ANCHOVETA
Materia dc.subject Ecosistemas marinos
Materia dc.subject REDES NEURONALES
Title dc.title Predicción de captura de anchovetas utilizando redes neuronales
Tipo dc.type texto


Archivos en el ítem

Archivos Tamaño Formato Ver

No hay archivos asociados a este ítem.

Este ítem aparece en la(s) siguiente(s) colección(ones)

Mostrar el registro sencillo del ítem