Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study
In recent years, machine-learning techniques, particularly deep learning, have outperformed traditional time-series forecasting approaches in many contexts, including univariate and multivariate predictions. This study aims to investigate the capability of (i) gated recurrent neural networks, includ...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-06-01
|
| Series: | Machine Learning with Applications |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666827022000275 |
| _version_ | 1818551949387104256 |
|---|---|
| author | Shahrokh Shahi Flavio H. Fenton Elizabeth M. Cherry |
| author_facet | Shahrokh Shahi Flavio H. Fenton Elizabeth M. Cherry |
| author_sort | Shahrokh Shahi |
| collection | DOAJ |
| description | In recent years, machine-learning techniques, particularly deep learning, have outperformed traditional time-series forecasting approaches in many contexts, including univariate and multivariate predictions. This study aims to investigate the capability of (i) gated recurrent neural networks, including long short-term memory (LSTM) and gated recurrent unit (GRU) networks, (ii) reservoir computing (RC) techniques, such as echo state networks (ESNs) and hybrid physics-informed ESNs, and (iii) the nonlinear vector autoregression (NVAR) approach, which has recently been introduced as the next generation RC, for the prediction of chaotic time series and to compare their performance in terms of accuracy, efficiency, and robustness. We apply the methods to predict time series obtained from two widely used chaotic benchmarks, the Mackey–Glass and Lorenz-63 models, as well as two other chaotic datasets representing a bursting neuron and the dynamics of the El Niño Southern Oscillation, and to one experimental dataset representing a time series of cardiac voltage with complex dynamics. We find that even though gated RNN techniques have been successful in forecasting time series generally, they can fall short in predicting chaotic time series for the methods, datasets, and ranges of hyperparameter values considered here. In contrast, for the chaotic datasets studied, we found that reservoir computing and NVAR techniques are more computationally efficient and offer more promise in long-term prediction of chaotic time series. |
| first_indexed | 2024-12-12T09:06:42Z |
| format | Article |
| id | doaj.art-92cd80f19028448e90d07de820105f1b |
| institution | Directory Open Access Journal |
| issn | 2666-8270 |
| language | English |
| last_indexed | 2024-12-12T09:06:42Z |
| publishDate | 2022-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Machine Learning with Applications |
| spelling | doaj.art-92cd80f19028448e90d07de820105f1b2022-12-22T00:29:39ZengElsevierMachine Learning with Applications2666-82702022-06-018100300Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative studyShahrokh Shahi0Flavio H. Fenton1Elizabeth M. Cherry2School of Computational Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332, United States of America; Corresponding author.School of Physics, Georgia Institute of Technology, Atlanta, GA 30332, United States of AmericaSchool of Computational Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332, United States of AmericaIn recent years, machine-learning techniques, particularly deep learning, have outperformed traditional time-series forecasting approaches in many contexts, including univariate and multivariate predictions. This study aims to investigate the capability of (i) gated recurrent neural networks, including long short-term memory (LSTM) and gated recurrent unit (GRU) networks, (ii) reservoir computing (RC) techniques, such as echo state networks (ESNs) and hybrid physics-informed ESNs, and (iii) the nonlinear vector autoregression (NVAR) approach, which has recently been introduced as the next generation RC, for the prediction of chaotic time series and to compare their performance in terms of accuracy, efficiency, and robustness. We apply the methods to predict time series obtained from two widely used chaotic benchmarks, the Mackey–Glass and Lorenz-63 models, as well as two other chaotic datasets representing a bursting neuron and the dynamics of the El Niño Southern Oscillation, and to one experimental dataset representing a time series of cardiac voltage with complex dynamics. We find that even though gated RNN techniques have been successful in forecasting time series generally, they can fall short in predicting chaotic time series for the methods, datasets, and ranges of hyperparameter values considered here. In contrast, for the chaotic datasets studied, we found that reservoir computing and NVAR techniques are more computationally efficient and offer more promise in long-term prediction of chaotic time series.http://www.sciencedirect.com/science/article/pii/S2666827022000275Recurrent neural networksReservoir computingEcho state networksDeep learningChaotic time seriesNonlinear vector autoregression |
| spellingShingle | Shahrokh Shahi Flavio H. Fenton Elizabeth M. Cherry Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study Machine Learning with Applications Recurrent neural networks Reservoir computing Echo state networks Deep learning Chaotic time series Nonlinear vector autoregression |
| title | Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study |
| title_full | Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study |
| title_fullStr | Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study |
| title_full_unstemmed | Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study |
| title_short | Prediction of chaotic time series using recurrent neural networks and reservoir computing techniques: A comparative study |
| title_sort | prediction of chaotic time series using recurrent neural networks and reservoir computing techniques a comparative study |
| topic | Recurrent neural networks Reservoir computing Echo state networks Deep learning Chaotic time series Nonlinear vector autoregression |
| url | http://www.sciencedirect.com/science/article/pii/S2666827022000275 |
| work_keys_str_mv | AT shahrokhshahi predictionofchaotictimeseriesusingrecurrentneuralnetworksandreservoircomputingtechniquesacomparativestudy AT flaviohfenton predictionofchaotictimeseriesusingrecurrentneuralnetworksandreservoircomputingtechniquesacomparativestudy AT elizabethmcherry predictionofchaotictimeseriesusingrecurrentneuralnetworksandreservoircomputingtechniquesacomparativestudy |