Research Article
BibTex RIS Cite

NARX Sinir Ağları Yöntemiyle 25. Güneş Aktivite Çevrimi Tahmini

Year 2022, , 57 - 60, 31.12.2022
https://doi.org/10.55064/tjaa.1037256

Abstract

Güneş Aktivite Çevrimlerini (GAC) tahmin etmek; Dünya yörüngesindeki uzay araçlarının güvenliği, iletişim ağları ve genel olarak yaşam üzerindeki olumsuz etkileri nedeniyle önemli hale gelmiştir. Bu çalışmada, 25. GAC tahmini için yaklaşık3246 adet 13 aylık ortalama Güneş Lekesi Sayısı’ndan (GLS) (Temmuz 1749 - Aralık 2019 arasında) oluşan büyük bir veri seti ile Doğrusal Olmayan Otoregresif Eksojen (NARX) sinir ağı tabanlı modeller kullanılmıştır. NARX modelleri veri setini Bayesian Regülasyonu kullanılarak eğitilmiştir. NARX modelinin performansı Ortalama Karekök Sapması (RMSE),Korelasyon Katsayısı (r) ve Nash-Sutcliffe Verimlilik Katsayısı (NSE) kullanılarak hesaplanmıştır ve modelin performansının“çok iyi” (NSE > 0.95) olduğu bulunmuştur. Modellerimiz diğer benzer çalışmalardan elde edilen sonuçlarla karşılaştırılarak doğrulanmıştır ve 25. GAC için maksimum GLS 104,542 ve maksimum ayı Mayıs 2024 olarak tahmin edilmiştir.NARX tabanlı modellerin literatürde bildirilen diğer yöntemlere kıyasla iyi ve uyumlu tahmin sonuçları ortaya çıkardığı görülmüştür.

References

  • Ahmed U., Mumtaz R., Anwar H., Shah A. A., Irfan R., García-NietoJ., 2019, Water, 11
  • Balogh A., Hudson H., Petrovay K., von Steiger R., 2014, SpaceScience Reviews, 186, 1
  • Boussaada Z., Curea O., Remaci A., Camblong H., Mrabet Bellaaj, N., 2018, Energies, 11
  • Du Z. L., 2020, Astrophys. Space Sci., 365(6), 134
  • Guzman S., Paz J., Tagert M., 2017, Water Resour Manage, 31,1591–1603
  • Han Y., Yin Z., 2019, Solar Physics, 294, 107
  • Hathaway D. H., 2015, Living Rev. Solar Phys., 12(1), 4
  • Helal R., Galal A., 2013, Journal of Advanced Research, 4, 275–278
  • Li F. Y., Kong D. F. Xie. J. L., Xiang N. B., C X. J., 2018, Journalof Atmospheric and Solar-Terrestrial Physics, 181, 110
  • Menezes J., Barreto G. D. A., 2006, 2006 Ninth Brazilian Symposiumon Neural Networks (SBRN’06), pp 160–165
  • Menezes J., Barreto G. D. A., 2008, Neurocomputing, 71, 3335
  • Petrovay K., 2020, Living Rev. Solar Phys., 17(1), 2
  • Sarp V., Kilcik A., 2018, Nonlinear Prediction of Solar Cycle 25
  • Sarp V., Kilcik A., Yurchyshyn V., Rozelot J. P., Özgüc A., 2018,Astrophys. Space Sci., 365(6), 2981
  • Svalgaard L., 2020, Prediction of Solar Cycle 25
  • Uwamahoro J., 2008, Master’s thesis, Rhodes University, core.ac.uk
  • Wu S. S., Qin G., 2021, Predicting Sunspot Numbers for Solar Cycles25 and 26 (arXiv:2102.06001)
  • Wunsch A., Liesch T., Broda S., 2018, Journal of Hydrology, 567
  • Yoshida A., 2014, Annales Geophysicae, 32, 1035
  • Yoshida A., Yamagishi H., 2010, Annales Geophysicae, 28, 417

Prediction of Solar Activity Cycle 25 with NARX Neural Networks

Year 2022, , 57 - 60, 31.12.2022
https://doi.org/10.55064/tjaa.1037256

Abstract

Negative effects on the stability of earth’s orbiting spacecrafts, telecommunication networks, and on life in general have been correlated with the solar activity cycles. This work is devoted for the development of a neural network-based models for the prediction and forecasting of forthcoming solar activity cycles. An approach is developed based on the nonlinear autoregressive exogenous (NARX) technique; we consider a long time series of observations. The prediction method is based on the nonlinear autoregressive network with exogenous inputs (NARX) with its ability to derive the underlying complex and nonlinear relationships. A big dataset of about 3246 monthly average sunspot numbers (SSN) (between Jul 1749 – December 2019) is used for the current study. The NARX model was trained with the data set using the Bayesian Regulation. The performance of the NARX model was evaluated using statistical parameters such as RMSE, r, and NSE and the performance of the model was found to be “very good” (NSE > 0.95). Our models were compared and verified with other similar studies and show that the current predicted maximum SSN for the Solar Cycle 25 is 104.542 and date of maxima is May 2024. We conclude that our currently NARX based approaches offer good and accurate prediction results in comparison to other methods reported in the literature.

References

  • Ahmed U., Mumtaz R., Anwar H., Shah A. A., Irfan R., García-NietoJ., 2019, Water, 11
  • Balogh A., Hudson H., Petrovay K., von Steiger R., 2014, SpaceScience Reviews, 186, 1
  • Boussaada Z., Curea O., Remaci A., Camblong H., Mrabet Bellaaj, N., 2018, Energies, 11
  • Du Z. L., 2020, Astrophys. Space Sci., 365(6), 134
  • Guzman S., Paz J., Tagert M., 2017, Water Resour Manage, 31,1591–1603
  • Han Y., Yin Z., 2019, Solar Physics, 294, 107
  • Hathaway D. H., 2015, Living Rev. Solar Phys., 12(1), 4
  • Helal R., Galal A., 2013, Journal of Advanced Research, 4, 275–278
  • Li F. Y., Kong D. F. Xie. J. L., Xiang N. B., C X. J., 2018, Journalof Atmospheric and Solar-Terrestrial Physics, 181, 110
  • Menezes J., Barreto G. D. A., 2006, 2006 Ninth Brazilian Symposiumon Neural Networks (SBRN’06), pp 160–165
  • Menezes J., Barreto G. D. A., 2008, Neurocomputing, 71, 3335
  • Petrovay K., 2020, Living Rev. Solar Phys., 17(1), 2
  • Sarp V., Kilcik A., 2018, Nonlinear Prediction of Solar Cycle 25
  • Sarp V., Kilcik A., Yurchyshyn V., Rozelot J. P., Özgüc A., 2018,Astrophys. Space Sci., 365(6), 2981
  • Svalgaard L., 2020, Prediction of Solar Cycle 25
  • Uwamahoro J., 2008, Master’s thesis, Rhodes University, core.ac.uk
  • Wu S. S., Qin G., 2021, Predicting Sunspot Numbers for Solar Cycles25 and 26 (arXiv:2102.06001)
  • Wunsch A., Liesch T., Broda S., 2018, Journal of Hydrology, 567
  • Yoshida A., 2014, Annales Geophysicae, 32, 1035
  • Yoshida A., Yamagishi H., 2010, Annales Geophysicae, 28, 417
There are 20 citations in total.

Details

Primary Language Turkish
Subjects Astronomical Sciences (Other)
Journal Section Articles
Authors

Mirkan Yusuf Kalkan 0000-0002-2077-7269

Ahmet Talât Saygaç 0000-0002-8331-7454

Diaa Fawzy This is me 0000-0003-1993-0681

Publication Date December 31, 2022
Submission Date December 16, 2021
Acceptance Date February 8, 2022
Published in Issue Year 2022

Cite

TJAA, Türk Astronomi Derneğinin (TAD) bir yayınıdır.