Abstract
Predicting the time series of 10.7-cm solar radio flux is a challenging task because of its daily variability. This paper proposed a non-linear method, a convolutional and recurrent neural network combined model to achieve end-to-end F10.7 forecasts. The network consists of a one-dimensional convolutional neural network and a long short-term memory network. The CNN network extracted features from F10.7 original data, then trained the feature signals in the long short-term memory network, and outputted the predicted values. The F10.7 daily data during 2003–2014 are used for the testing set. The mean absolute percentage error values of approximately 2.04%, 2.78%, and 4.66% for 1-day, 3-day, and 7-day forecasts, respectively. The statistical results of evaluating the root mean square error, spearman correlation coefficient shows a superior effect as a whole for the 1–27 days forecast, compared with the ordinary single neural network and combination models.
Highlights
The results showed that the convolutional neural network (CNN)-long short-term memory network (LSTM) model had Mean Absolute Percentage Error (MAPE) values of 2.04% (1-day ahead), 2.78%
The CNN-LSTM model was found to be effective in improving the prediction accuracy of high-solar-activity years
The CNN-LSTM was compared with the Y model and the CH-MF models, and showed distinct advantages in a variety of evaluation metrics
Summary
The 10.7 cm solar radio flux is the solar radio emission intensity in a 100-MHz-wide band centered at 2800 MHz (i.e., 10.7-cm wavelength). Most studies in F10.7 forecasting have only focused on constructing linear models, which usually have a relatively stable effect in mid- and long-term prediction. [13], a support vector machine (SVR) was used to make a short-term prediction of F10.7 This method reduced the computational complexity in the training process through a learning-based algorithm and achieves an accuracy close to that of traditional feed-forward neural network with fewer data. We proposed a time series prediction model for the short- and mid-term F10.7 forecasting to address the above deficiencies. The model is comprised two independent networks: the feature extraction network and the predicting network, respectively devised by a onedimension convolutional neural network (CNN) and a long short-term memory network (LSTM).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
