A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity
The accurate long-term forecasting of solar activity is crucial in the current era of space explorations and in the study of planetary climate evolution. With timescales of about 11 years, these forecasts deal with the prediction of the very general features of a solar cycle such as its amplitude, p...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Engineering Proceedings |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2673-4591/39/1/16 |
| _version_ | 1797580296316518400 |
|---|---|
| author | Adriana Marcucci Giovanna Jerse Valentina Alberti Mauro Messerotti |
| author_facet | Adriana Marcucci Giovanna Jerse Valentina Alberti Mauro Messerotti |
| author_sort | Adriana Marcucci |
| collection | DOAJ |
| description | The accurate long-term forecasting of solar activity is crucial in the current era of space explorations and in the study of planetary climate evolution. With timescales of about 11 years, these forecasts deal with the prediction of the very general features of a solar cycle such as its amplitude, peak time and period. Solar radio indices, continuously measured by a network of ground-based solar radio telescopes, are among the most commonly used descriptors to characterise the solar activity level. They can act as proxies for the strength of ionising radiations, such as solar ultraviolet and X-ray emissions, which directly affect the atmospheric density. In a preliminary comparative study of a selection of univariate deep-learning methods targeting medium-term forecasts of the F10.7 index, we noticed that the performance of all the considered models tends to degrade with increasing timescales and that this effect is smoother when a multi-attention module is included in the used neural network architecture. In this work, we present a multivariate approach based on the combination of fast iterative filtering (FIF) algorithm, long-short term memory (LSTM) network and multi-attention module, trained for the present solar cycle forecasting. Several solar radio flux time series, namely F3.2, F8, F10.7, F15, F30, are fed into the neural network to forecast the F10.7 index. The results are compared with the official solar cycle forecasting released by the Solar Cycle Prediction Panel representing NOAA, NASA and the International Space Environmental Services (ISES) to highlight possible discrepancies. |
| first_indexed | 2024-03-10T22:48:06Z |
| format | Article |
| id | doaj.art-38bf29c7aef449aca48c8c49f6748731 |
| institution | Directory Open Access Journal |
| issn | 2673-4591 |
| language | English |
| last_indexed | 2024-03-10T22:48:06Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Engineering Proceedings |
| spelling | doaj.art-38bf29c7aef449aca48c8c49f67487312023-11-19T10:30:27ZengMDPI AGEngineering Proceedings2673-45912023-06-013911610.3390/engproc2023039016A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar ActivityAdriana Marcucci0Giovanna Jerse1Valentina Alberti2Mauro Messerotti3Department of Physics, University of Trieste, Via A. Valerio 2, 34127 Trieste, ItalyAstronomical Observatory of Trieste, INAF, Via G. Tiepolo 11, 34143 Trieste, ItalyAstronomical Observatory of Trieste, INAF, Via G. Tiepolo 11, 34143 Trieste, ItalyAstronomical Observatory of Trieste, INAF, Via G. Tiepolo 11, 34143 Trieste, ItalyThe accurate long-term forecasting of solar activity is crucial in the current era of space explorations and in the study of planetary climate evolution. With timescales of about 11 years, these forecasts deal with the prediction of the very general features of a solar cycle such as its amplitude, peak time and period. Solar radio indices, continuously measured by a network of ground-based solar radio telescopes, are among the most commonly used descriptors to characterise the solar activity level. They can act as proxies for the strength of ionising radiations, such as solar ultraviolet and X-ray emissions, which directly affect the atmospheric density. In a preliminary comparative study of a selection of univariate deep-learning methods targeting medium-term forecasts of the F10.7 index, we noticed that the performance of all the considered models tends to degrade with increasing timescales and that this effect is smoother when a multi-attention module is included in the used neural network architecture. In this work, we present a multivariate approach based on the combination of fast iterative filtering (FIF) algorithm, long-short term memory (LSTM) network and multi-attention module, trained for the present solar cycle forecasting. Several solar radio flux time series, namely F3.2, F8, F10.7, F15, F30, are fed into the neural network to forecast the F10.7 index. The results are compared with the official solar cycle forecasting released by the Solar Cycle Prediction Panel representing NOAA, NASA and the International Space Environmental Services (ISES) to highlight possible discrepancies.https://www.mdpi.com/2673-4591/39/1/16solar activity forecastingsolar radio indexdeep learningmultivariate predictiontime series forecastingmulti-attention |
| spellingShingle | Adriana Marcucci Giovanna Jerse Valentina Alberti Mauro Messerotti A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity Engineering Proceedings solar activity forecasting solar radio index deep learning multivariate prediction time series forecasting multi-attention |
| title | A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity |
| title_full | A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity |
| title_fullStr | A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity |
| title_full_unstemmed | A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity |
| title_short | A Deep Learning Model Based on Multi-Head Attention for Long-Term Forecasting of Solar Activity |
| title_sort | deep learning model based on multi head attention for long term forecasting of solar activity |
| topic | solar activity forecasting solar radio index deep learning multivariate prediction time series forecasting multi-attention |
| url | https://www.mdpi.com/2673-4591/39/1/16 |
| work_keys_str_mv | AT adrianamarcucci adeeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT giovannajerse adeeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT valentinaalberti adeeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT mauromesserotti adeeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT adrianamarcucci deeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT giovannajerse deeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT valentinaalberti deeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity AT mauromesserotti deeplearningmodelbasedonmultiheadattentionforlongtermforecastingofsolaractivity |