Abstract
We introduce a data-centric framework that forecasts the four key solar drivers, F10.7, S10.7, M10.7, and Y10.7, with PatchTST, a multivariate time-series Transformer that reads an 18-day look-back window and delivers 6-day predictions, capturing short- and long-range patterns in one pass. To offset the natural bias toward quiet Sun conditions, we design a class-balanced loss that weights each sample by its distance from the historical activity distribution, giving storm-level episodes their fair share of influence. Learning is further stabilized with Reversible Instance Normalization, channel-independent embeddings, and an ensemble of weighted MAE and MSE objectives. Tested on the public SET benchmark, the model cuts mean percentage error by 77.7 %and standard percentage deviation by 60.2 % versus SOLAR2000, with gains of more than 80 % during solar-storm periods, and shows less than 2.3 percentage-point drift on a distribution-matched validation set. Training fƒinishes in under eight minutes on a single RTX 3090 and inference stays below two seconds, making the approach viable for near-real-time space-weather forecasting and safer satellite operations.
Abstract
We introduce a data-centric framework that forecasts the four key solar drivers, F10.7, S10.7, M10.7, and Y10.7, with PatchTST, a multivariate time-series Transformer that reads an 18-day look-back window and delivers 6-day predictions, capturing short- and long-range patterns in one pass. To offset the natural bias toward quiet Sun conditions, we design a class-balanced loss that weights each sample by its distance from the historical activity distribution, giving storm-level episodes their fair share of influence. Learning is further stabilized with Reversible Instance Normalization, channel-independent embeddings, and an ensemble of weighted MAE and MSE objectives. Tested on the public SET benchmark, the model cuts mean percentage error by 77.7 %and standard percentage deviation by 60.2 % versus SOLAR2000, with gains of more than 80 % during solar-storm periods, and shows less than 2.3 percentage-point drift on a distribution-matched validation set. Training fƒinishes in under eight minutes on a single RTX 3090 and inference stays below two seconds, making the approach viable for near-real-time space-weather forecasting and safer satellite operations. Read More
Related posts:
Over 200,000 new HGV drivers needed by 2029, RHA forecasts
Nonparametric data segmentation in multivariate time series via joint characteristic functions
FedEx expert shares advice for drivers as Storm Éowyn hits UK and Ireland
Treatment of Solar Radiation by Spatial and Temporal Discretization for Modeling the Thermal Response of Arch Dams
Environmental factors determining weed species distribution in organic manure and inorganic fertiliser smallholder maize (Zea mays L.) production
An Operational Framework for Land Cover Classification in the Context of REDD+ Mechanisms. A Case Study from Costa Rica