Sunspot Forecasting by Using Chaotic Time-series Analysis and NARX Network

  title={Sunspot Forecasting by Using Chaotic Time-series Analysis and NARX Network},
  author={Chuanjin Jiang and Fugen Song},
  journal={J. Comput.},
Chaotic time-series is a dynamic nonlinear system whose features can not be fully reflected by Linear Regression Model or Static Neural Network. While Nonlinear Autoregressive with eXogenous input includes feedback of network output, therefore, it can better reflect the system’s dynamic feature. Take annual active times of sunspot as an example, after verifying the chaos of sunspot time-series and calculating the series’ embedding dimension and delay, we establish sunspot prediction model with… 

Figures and Tables from this paper

Investigation of Chaotic Nature of Sunspot Data by Nonlinear Analysis Techniques
In this work, an attempt is made to detect the chaotic nature of smooth monthly sunspot (SSN) time series using various nonlinear analysis techniques to quantify the uncertainty involved. Different
Time Series Forecasting with Trend and Seasonal Patterns using NARX Network Ensembles
The empirical study results show that the NARX model with binary dummy exogenous is the most accurate for trend and seasonal with multiplicative properties data patterns and the median and mode ensemble operators, which are rarely used, are more accurate than the mean ensemble operator for data that have Trend and seasonal patterns.
Forecasting of Sunspot Time Series Using a Hybridization of ARIMA, ETS and SVM Methods
A hybridization of the autoregressive integrated moving average (ARIMA); exponential smoothing with error, trend and seasonality (ETS); and support vector machine (SVM) to predict monthly and yearly SN time series is proposed.
A hybrid elman neural network predictor for time series prediction
The results confirm that the proposed Elman network hybridised with time delay inputs can predict the time series with more accurately and effectively than the existing methods.
A comparison of learning algorithms for seasonal time series forecasting using NARX model
The results of the case study show that the best approach to the NARX model is to use binary dummy variables and data with the first differencing process, and the GRPROP algorithm shows the least computation time, the fastest training process steps, and forecasting accuracy with the best MAPE value.
Weather prediction by recurrent neural network dynamics
A multiple weather attributes forecasting method for one-day-ahead prediction using NARX NN in local scale using collected historical records of weather station from 1980 to 2009 is presented.
Neural network forecast of the sunspot diagram
The analysis of the results show that it is indeed possible to reconstruct the overall shape and amplitude of the spatial-temporal pattern of sunspots using these feed-forward neural networks, but it is concluded that more data and/or improved neural network techniques are probably necessary for this approach to have real predictive power.
Dynamic Neural Network Based Very Short-Term Wind Speed Forecasting
The result shows that the proposed model outperforms the BPNN based on the metrics used, and the nonlinear autoregressive model with exogenous inputs (NARX) is proposed for wind speed forecast.
Sunspots Time-Series Prediction Based on Complementary Ensemble Empirical Mode Decomposition and Wavelet Neural Network
The proposed prediction model based on complementary ensemble empirical mode decomposition (CEEMD) and wavelet neural network (WNN) is proposed to improve the prediction accuracy of sunspot time series and has better prediction accuracy and smaller error.
Online multistep-ahead inundation depth forecasts by recurrent NARX networks
It is demonstrated that the R-NARX model can effectively inhibit error growth and accumulation when being applied to online multistep-ahead inundation forecasts over a long lasting forecast period.


Modeling and Multi-Step Prediction of Chaotic Time Series Based on RBF Neural Networks
Simulations show that RBF networks models have good fitness and high accuracy of single and multistep prediction to the chaotic time series.
A delay damage model selection algorithm for NARX neural networks
It is shown that using intelligent memory order selection through pruning and good initial heuristics significantly improves the generalization and predictive performance of these nonlinear systems on problems as diverse as grammatical inference and time series prediction.
On the numerical determination of the dimension of an attractor
Input–output parametric models for nonlinear systems
  • International Journal of Control,
  • 1985
Much A do about Nothing? Exchange Rates Forecasting: Neural Networks vs. Linear Models
  • Using Monthly and Weekly Data”,
  • 1996
The Modeling and Forecasting of Chaotic Time Series Based on Neural Network
  • Comuputer Engineering and Applictions,
  • 2005