Time series prediction using artificial neural networks
In this work, artificial neural network (ANN) algorithms were used to predict time series of the oceanographic variables Southern Oscilation Index (SOI) and sea surface temperature anomaly (SSTA). The finite impulse response neural network (FIRNN) was applied to data obtained from the NOAA...
محفوظ في:
المؤلفون الرئيسيون: | , , |
---|---|
التنسيق: | Online |
اللغة: | eng |
منشور في: |
Iniversidad Autónoma de Baja California
2002
|
الوصول للمادة أونلاين: | https://www.cienciasmarinas.com.mx/index.php/cmarinas/article/view/205 |
الوسوم: |
إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
|
id |
oai:cienciasmarinas.com.mx:article-205 |
---|---|
record_format |
ojs |
spelling |
oai:cienciasmarinas.com.mx:article-2052019-05-02T21:16:13Z Time series prediction using artificial neural networks Predicción de series de tiempo aplicando redes neuronales artificiales Pérez-Chavarría, MA Hidalgo-Silva, HH Ocampo-Torres, FJ forecasting time series neural-network predicción serie-temporal red-neurona In this work, artificial neural network (ANN) algorithms were used to predict time series of the oceanographic variables Southern Oscilation Index (SOI) and sea surface temperature anomaly (SSTA). The finite impulse response neural network (FIRNN) was applied to data obtained from the NOAA. In order to determine the most efficient FIRNN architecture, several experiments were made varying different parameters. The best predictions were obtained for a network with one input neuron and 10th-order filters in the input layer, two 8-neuron 5th-order filter hidden layers and one output neuron. All the networks were trained with the temporal backpropagation learning algorithm, using the sigmoid transfer function at the hidden layers and a linear output. The learning rate was 0.001. In most experiments a normalized mean square error of 0.4 ± 0.1 and a correlation coefficient between the original and the predicted series greater than 0.8 were found. From a comparison with other SSTA prediction methods, the results obtained with the neural network were the best ones for the short term forecasting case. En este trabajo se presentan algoritmos basados en redes neuronales artificiales (RNA) para la predicción de series temporales de las variables oceanográficas Índice de Oscilación del Sur (IOS) y anomalías de temperatura superficial del mar (ATSM). Se realizaron experimentos de predicción utilizando la arquitectura de redes neuronales conocida como redes con respuesta finita al impulso (RNRFI). Se hicieron experimentos variando los diferentes parámetros de la RNRFI, para determinar aquéllos que permitieran un mejor comportamiento. Se encontró que para series temporales del IOS y las ATSMs los mejores resultados se presentan al conformar una RNRFI con una neurona de entrada, filtros de orden 10 en la capa de entrada, dos capas posteriores de 8 neuronas con filtros de orden 5 para cada una de ellas, y una neurona de salida. Todas las redes fueron entrenadas con el algoritmo de aprendizaje retropropagación temporal, usando la sigmoide como función de activación en las capas ocultas y salida lineal. La razón de aprendizaje usada fue de 0.001. En la mayoría de los experimentos realizados se obtuvo un error cuadrático medio normalizado de 0.4 ± 0.1 y un coeficiente de correlación mayor que 0.8 entre la serie original y la predicha. Para el caso de las ATSMs, se observó que las RNA tuvieron un mejor comportamiento que otros métodos de predicción, considerando predicciones a corto plazo. Iniversidad Autónoma de Baja California 2002-03-06 info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion Peer-reviewed Article Artículo Arbitrado application/pdf https://www.cienciasmarinas.com.mx/index.php/cmarinas/article/view/205 10.7773/cm.v28i1.205 Ciencias Marinas; Vol. 28 No. 1 (2002); 67-77 Ciencias Marinas; Vol. 28 Núm. 1 (2002); 67-77 2395-9053 0185-3880 eng https://www.cienciasmarinas.com.mx/index.php/cmarinas/article/view/205/1185 |
institution |
Ciencias Marinas |
collection |
OJS |
language |
eng |
format |
Online |
author |
Pérez-Chavarría, MA Hidalgo-Silva, HH Ocampo-Torres, FJ |
spellingShingle |
Pérez-Chavarría, MA Hidalgo-Silva, HH Ocampo-Torres, FJ Time series prediction using artificial neural networks |
author_facet |
Pérez-Chavarría, MA Hidalgo-Silva, HH Ocampo-Torres, FJ |
author_sort |
Pérez-Chavarría, MA |
title |
Time series prediction using artificial neural networks |
title_short |
Time series prediction using artificial neural networks |
title_full |
Time series prediction using artificial neural networks |
title_fullStr |
Time series prediction using artificial neural networks |
title_full_unstemmed |
Time series prediction using artificial neural networks |
title_sort |
time series prediction using artificial neural networks |
description |
In this work, artificial neural network (ANN) algorithms were used to predict time series of the oceanographic variables Southern Oscilation Index (SOI) and sea surface temperature anomaly (SSTA). The finite impulse response neural network (FIRNN) was applied to data obtained from the NOAA. In order to determine the most efficient FIRNN architecture, several experiments were made varying different parameters. The best predictions were obtained for a network with one input neuron and 10th-order filters in the input layer, two 8-neuron 5th-order filter hidden layers and one output neuron. All the networks were trained with the temporal backpropagation learning algorithm, using the sigmoid transfer function at the hidden layers and a linear output. The learning rate was 0.001. In most experiments a normalized mean square error of 0.4 ± 0.1 and a correlation coefficient between the original and the predicted series greater than 0.8 were found. From a comparison with other SSTA prediction methods, the results obtained with the neural network were the best ones for the short term forecasting case. |
publisher |
Iniversidad Autónoma de Baja California |
publishDate |
2002 |
url |
https://www.cienciasmarinas.com.mx/index.php/cmarinas/article/view/205 |
_version_ |
1715723941976211456 |