An End-to-end Ensemble Machine Learning Approach for Predicting High-impact Solar Energetic Particle Events Using Multimodal Data

The Astrophysical Journal Supplement Series, Journal Year: 2025, Volume and Issue: 277(2), P. 34 - 34

Published: March 17, 2025

Abstract Solar energetic particle (SEP) events, in particular high-energy-range SEP pose significant risks to space missions, astronauts, and technological infrastructure. Accurate prediction of these high-impact events is crucial for mitigating potential hazards. In this study, we present an end-to-end ensemble machine learning (ML) framework the ∼100 MeV events. Our approach leverages diverse data modalities sourced from Heliospheric Observatory Geostationary Operational Environmental Satellite integrating extracted active region polygons solar extreme ultraviolet (EUV) imagery, time-series proton flux measurements, sunspot activity data, detailed characteristics. To quantify predictive contribution each modality (e.g., EUV or time series), independently evaluate them using a range ML models assess their performance forecasting Finally, enhance performance, train model that combines all trained on individual modalities, leveraging strengths modality. proposed shows promising achieving recall 0.80 0.75 balanced imbalanced settings, respectively, underscoring effectiveness multimodal integration robust event enhanced capabilities.

Language: Английский

---

Multi‐Source Forecast of Solar Cycle Flare Activity Using the Novel Informer‐Based Models

Space Weather, Journal Year: 2025, Volume and Issue: 23(4)

Published: April 1, 2025

Abstract Solar flares, the significant indicators of solar activity, have an impact on Earth's satellites and communication systems. Accurate prediction flare events is crucial for mitigating these effects. In this work, we use multiple data sources, including Geostationary Operational Environmental Satellites soft X‐ray flux index, to forecast activity during Cycle 25 (SC25). Our results show that: (a) The north‐south asymmetry SC25 well revealed, southern hemisphere greater than northern one. (b) Gnevyshev peaks chromospheric are clearly identified they deeper other atmospheric indicators. different timescales responses geomagnetic interplanetary magnetic fields may be cause peaks. (c) Chromospheric lags behind photospheric sunspot indicating that changes in precede events. (d) level influenced by modulating effect Gleissberg Cycle, as supported precursor indices. These offer valuable insights into temporal spatial distribution SC25.

Language: Английский

---

Investigating Performance Trends of Simulated Real-time Solar Flare Predictions: The Impacts of Training Windows, Data Volumes, and the Solar Cycle

The Astrophysical Journal, Journal Year: 2024, Volume and Issue: 964(2), P. 163 - 163

Published: March 28, 2024

Abstract This study explores the behavior of machine-learning-based flare forecasting models deployed in a simulated operational environment. Using Georgia State University’s Space Weather Analytics for Solar Flares benchmark data set, we examine impacts training methodology and solar cycle on decision tree, support vector machine, multilayer perceptron performance. We implement our classifiers using three temporal windows: stationary, rolling, expanding. The stationary window trains single set available before first instance, which remains constant throughout cycle. rolling from time interval moves with Finally, expanding all instance. For each window, number input features (1, 5, 10, 25, 50, 120) sizes (5, 8, 11, 14, 17, 20 months) were tested. To surprise, found that, months, skill scores comparable regardless type, feature count, classifier selected. Furthermore, reducing size this only marginally decreased implies given enough data, can be chosen over other types, eliminating need model retraining. moderately strong positive correlation was to exist between model’s false-positive rate X-ray background flux. suggests that phase has considerable influence forecasting.

Language: Английский

---

Time Series of Magnetic Field Parameters of Merged MDI and HMI Space-weather Active Region Patches as Potential Tool for Solar Flare Forecasting

The Astrophysical Journal, Journal Year: 2024, Volume and Issue: 972(2), P. 169 - 169

Published: Sept. 1, 2024

Abstract Solar flare prediction studies have been recently conducted with the use of Space-Weather MDI (Michelson Doppler Imager on board and Heliospheric Observatory) Active Region Patches (SMARPs) HMI (Helioseismic Magnetic Dynamics (SHARPs), which are two currently available data products containing magnetic field characteristics solar active regions (ARs). The present work is an effort to combine them into one product, perform some initial statistical analyses in order further expand their application space-weather forecasting. combined derived by filtering, rescaling, merging SMARP SHARP parameters, can then be spatially reduced create uniform multivariate time series. resulting MDI–HMI set spans period between 1996 April 4 2022 December 13, may extended a more recent date. This provides opportunity correlate compare it other series, such as daily index or properties soft X-ray flux measured Geostationary Operational Environmental Satellites. Time-lagged cross correlation indicates that relationship exist, where ARs lead time. Applying rolling-window technique makes possible see how this leader–follower dynamic varies Preliminary results indicate areas high generally correspond increased activity during peak cycle.

Language: Английский

---

The Random Hivemind: An ensemble deep learning application to the solar energetic particle prediction problem

Advances in Space Research, Journal Year: 2024, Volume and Issue: 74(12), P. 6252 - 6263

Published: April 27, 2024

Language: Английский

---

The Mars Mission Data from Curiosity Rover Measuring Radiation Exposure

Published: Jan. 1, 2024

Language: Английский

---