Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method
As an important index of solar activity, the 10.7-cm solar radio flux (<i>F</i><sub>10.7</sub>) can indicate changes in the solar EUV radiation, which plays an important role in the relationship between the Sun and the Earth. Therefore, it is valuable to study and forecast &l...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-01-01
|
| Series: | Universe |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2218-1997/8/1/30 |
| _version_ | 1797489940472266752 |
|---|---|
| author | Wanting Zhang Xinhua Zhao Xueshang Feng Cheng’ao Liu Nanbin Xiang Zheng Li Wei Lu |
| author_facet | Wanting Zhang Xinhua Zhao Xueshang Feng Cheng’ao Liu Nanbin Xiang Zheng Li Wei Lu |
| author_sort | Wanting Zhang |
| collection | DOAJ |
| description | As an important index of solar activity, the 10.7-cm solar radio flux (<i>F</i><sub>10.7</sub>) can indicate changes in the solar EUV radiation, which plays an important role in the relationship between the Sun and the Earth. Therefore, it is valuable to study and forecast <i>F</i><sub>10.7</sub>. In this study, the long short-term memory (LSTM) method in machine learning is used to predict the daily value of <i>F</i><sub>10.7</sub>. The <i>F</i><sub>10.7</sub> series from 1947 to 2019 are used. Among them, the data during 1947–1995 are adopted as the training dataset, and the data during 1996–2019 (solar cycles 23 and 24) are adopted as the test dataset. The fourfold cross validation method is used to group the training set for multiple validations. We find that the root mean square error (RMSE) of the prediction results is only 6.20~6.35 sfu, and the correlation coefficient (R) is as high as 0.9883~0.9889. The overall prediction accuracy of the LSTM method is equivalent to those of the widely used autoregressive (AR) and backpropagation neural network (BP) models. Especially for 2-day and 3-day forecasts, the LSTM model is slightly better. All this demonstrates the potentiality of the LSTM method in the real-time forecasting of <i>F</i><sub>10.7</sub> in future. |
| first_indexed | 2024-03-10T00:23:49Z |
| format | Article |
| id | doaj.art-7680036a9d494016be405c2d5f378d77 |
| institution | Directory Open Access Journal |
| issn | 2218-1997 |
| language | English |
| last_indexed | 2024-03-10T00:23:49Z |
| publishDate | 2022-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Universe |
| spelling | doaj.art-7680036a9d494016be405c2d5f378d772023-11-23T15:37:13ZengMDPI AGUniverse2218-19972022-01-01813010.3390/universe8010030Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory MethodWanting Zhang0Xinhua Zhao1Xueshang Feng2Cheng’ao Liu3Nanbin Xiang4Zheng Li5Wei Lu6State Key Laboratory of Space Weather, National Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaState Key Laboratory of Space Weather, National Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaState Key Laboratory of Space Weather, National Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaCAEIT, Beijing 100041, ChinaYunnan Observatories, Chinese Academy of Sciences, Kunming 650011, ChinaInstitute of Space Weather, Nanjing University of Information Science & Technology, Nanjing 210044, ChinaState Key Laboratory of Space Weather, National Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaAs an important index of solar activity, the 10.7-cm solar radio flux (<i>F</i><sub>10.7</sub>) can indicate changes in the solar EUV radiation, which plays an important role in the relationship between the Sun and the Earth. Therefore, it is valuable to study and forecast <i>F</i><sub>10.7</sub>. In this study, the long short-term memory (LSTM) method in machine learning is used to predict the daily value of <i>F</i><sub>10.7</sub>. The <i>F</i><sub>10.7</sub> series from 1947 to 2019 are used. Among them, the data during 1947–1995 are adopted as the training dataset, and the data during 1996–2019 (solar cycles 23 and 24) are adopted as the test dataset. The fourfold cross validation method is used to group the training set for multiple validations. We find that the root mean square error (RMSE) of the prediction results is only 6.20~6.35 sfu, and the correlation coefficient (R) is as high as 0.9883~0.9889. The overall prediction accuracy of the LSTM method is equivalent to those of the widely used autoregressive (AR) and backpropagation neural network (BP) models. Especially for 2-day and 3-day forecasts, the LSTM model is slightly better. All this demonstrates the potentiality of the LSTM method in the real-time forecasting of <i>F</i><sub>10.7</sub> in future.https://www.mdpi.com/2218-1997/8/1/30solar radio fluxtime series forecastlong short-term memory |
| spellingShingle | Wanting Zhang Xinhua Zhao Xueshang Feng Cheng’ao Liu Nanbin Xiang Zheng Li Wei Lu Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method Universe solar radio flux time series forecast long short-term memory |
| title | Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method |
| title_full | Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method |
| title_fullStr | Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method |
| title_full_unstemmed | Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method |
| title_short | Predicting the Daily 10.7-cm Solar Radio Flux Using the Long Short-Term Memory Method |
| title_sort | predicting the daily 10 7 cm solar radio flux using the long short term memory method |
| topic | solar radio flux time series forecast long short-term memory |
| url | https://www.mdpi.com/2218-1997/8/1/30 |
| work_keys_str_mv | AT wantingzhang predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod AT xinhuazhao predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod AT xueshangfeng predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod AT chengaoliu predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod AT nanbinxiang predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod AT zhengli predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod AT weilu predictingthedaily107cmsolarradiofluxusingthelongshorttermmemorymethod |