Jurnal Internasional Linear Multi-Langkah F10.7 Peramalan Berdasarkan Korelasi Tugas dan Heteroskedastisitas – Wang – – Bumi dan Ilmu Luar Angkasa
Solar 10.7 cm radio flux (F10.7) is an important measure of solar radio emission activity. Estimated F10.7 Accurately plays a key role in space weather and forecasts of the global environment. We find that forecasting errors are heteroscedastic, something that is not often considered in previous models. In addition, task correlations between different forecasting steps are ignored in the current multi-step forecast model. In this work, we propose a linear multi-step forecasting model based on the correlation between the different forecasting steps and the characteristics of heteroscedasticity. Next, we introduce variational Bayesian procedures to optimize the model. The performance of the proposed model was tested on F10.7 historical data. The results show that the proposed model improves the performance of multi-step F10.7 estimates by considering correlation and heteroscedasticity.