A5080491475- Solar and Space Plasma Dynamics
- Solar Radiation and Photovoltaics
- Oil, Gas, and Environmental Issues
- Ionosphere and magnetosphere dynamics
- Market Dynamics and Volatility
- Tropical and Extratropical Cyclones Research
University of Chinese Academy of Sciences
2022-2023
National Space Science Center
2022-2023
Chinese Academy of Sciences
2022-2023
Solar flare forecasting is an essential component in space environment forecasting. Most of the deep learning models constructed are based on magnetograms active regions. Affected by projection effect, these can only forecast region center Sun. It difficult to meet need operational solar full disk. Based traditional activity parameters, this study, relationships between magnetic type region, area history outburst, 10 cm radio flux and flares from January 1996 December 2022 were statistically...
Abstract Recently, although various deep learning techniques have been applied to building space weather prediction models, a large amount of relevant prior knowledge solar eruptions and magnetic properties is ignored during the model development. By integrating in flare production into convolutional neural network (CNN) structures, we developed knowledge‐informed aiming at forecasting flares. The line‐of‐sight magnetograms Space‐weather HMI Active Region Patches (SHARP) from May 2010...
Solar flare forecasting is one of major components operational space weather forecasting. Complex active regions (ARs) are the main source producing flares, but only a few studies carried out to establish models for these ARs. In this study, four deep learning models, called Active Region Flare Forecasting Model (CARFFM)-1, −2, −3, and −4, established. They take AR longitudinal magnetic fields, vector fields total unsigned flux in neutral line region, region as input, respectively. These can...
The Spaceweather HMI Active Region Patch (SHARP) parameters have been widely used to develop flare prediction models. relatively small number of strong-flare events leads an unbalanced dataset that models can be sensitive the data and might lead bias limited performance. In this study, we adopted logistic regression algorithm a model for next 48 h based on SHARP parameters. was trained with five different inputs. first input original dataset; second third inputs were obtained by using two...