pith. sign in

arxiv: 2605.24748 · v1 · pith:X6N44XJInew · submitted 2026-05-23 · 🌌 astro-ph.SR · cs.LG

Deep Learning-Enabled Prediction of Geoeffective CMEs Using SOHO and SDO Observations

Zhaoxin Yan , Jason T. L. Wang , Haimin Wang , Harim Lee , Ju Jing , Yan Xu , Chunhui Xu , Vasyl Yurchyshyn This is my paper

Pith reviewed 2026-06-30 12:01 UTC · model grok-4.3

classification 🌌 astro-ph.SR cs.LG
keywords coronal mass ejectiongeomagnetic stormdeep learningfusion modelspace weatherSOHOSDOconvolutional neural network
0
0 comments X

The pith

A fusion deep learning model predicts whether Earth-directed coronal mass ejections will cause geomagnetic storms from SOHO and SDO images.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents a fusion model that applies convolutional neural networks to extract features from solar images and uses a prediction network to combine those features for classifying CME geoeffectiveness. The model trains on LASCO coronagraph data from SOHO together with AIA and HMI observations from SDO. Five-fold cross-validation yields a mean true skill statistic of 0.703 for deterministic forecasts and a mean Brier score of 0.095 for probabilistic outputs. These capabilities address the need to anticipate geomagnetic storms that can disrupt satellites and ground infrastructure.

Core claim

The fusion model integrates convolutional neural networks for learning features from multi-instrument solar observations with a prediction network that fuses the features and classifies whether an Earth-directed CME will produce a geomagnetic storm. Evaluated via five-fold cross validation on events observed by SOHO LASCO and SDO AIA/HMI, the model reaches a mean TSS of 0.703 when used deterministically and a mean Brier score of 0.095 when used probabilistically.

What carries the argument

The fusion model that combines convolutional neural network feature extraction from LASCO, AIA, and HMI images with a prediction network for feature fusion and binary or probabilistic classification of geoeffectiveness.

If this is right

  • The model supports both deterministic classification and probabilistic estimation of whether a CME will trigger a geomagnetic storm.
  • Cross-validation results indicate the approach can forecast the causal connection between Earth-directed CMEs and geomagnetic storms.
  • The reported TSS and Brier scores establish quantitative benchmarks for image-based geoeffectiveness prediction using SOHO and SDO data.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Operational deployment could shorten the warning time available for space weather impacts once real-time image pipelines are integrated.
  • The same fusion architecture might extend to forecasting other solar eruptive events such as flares if suitable labeled image sets exist.
  • Combining the model with in-situ solar wind measurements could test whether image-only skill improves when additional inputs are added.

Load-bearing premise

The geoeffectiveness labels assigned to the CME events are accurate and the selected events represent a typical distribution of Earth-directed CMEs.

What would settle it

Running the trained model on a fresh collection of Earth-directed CMEs whose geomagnetic storm outcomes have been independently verified and obtaining a TSS well below 0.7 would falsify the reported performance.

Figures

Figures reproduced from arXiv: 2605.24748 by Chunhui Xu, Haimin Wang, Harim Lee, Jason T. L. Wang, Ju Jing, Vasyl Yurchyshyn, Yan Xu, Zhaoxin Yan.

Figure 1
Figure 1. Figure 1: Distribution of the Dst index values for the 164 Earth-reaching CME events in our dataset [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Breakdown of the geoeffective (blue) and non-geoeffective (orange) CME events in our dataset. When training and validating our fusion model, we used solar observations from multiple instruments, including SOHO/LASCO, SDO/AIA with two chan￾nels 193 ˚A and 211 ˚A, and SDO/HMI [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Solar observations on the CME event that occurred on 14 March 2015 UT. Shown from left to right are a LASCO C2 image, an AIA 193 ˚A image, an AIA 211 ˚A image, and a full-disk HMI line-of-sight magnetogram. Wang 2019). Combining these observations gives a multimodal view of both the magnetic environment and the coronal response, improving the ability to assess CME potential and forecast their geoeffectiven… view at source ↗
Figure 4
Figure 4. Figure 4: Overall architecture of our fusion model. The model accepts, as input, a CME event represented by images of four types (LASCO C2, AIA 193 ˚A, AIA 211 ˚A and HMI), and assigns a pair of networks, ResNet (RN) and EfficientNet (EN), to learn and extract features from the input images of each type. These learned features are then sent to our PredictionNet (PN), which is composed of three modules: ensemble (EM)… view at source ↗
Figure 5
Figure 5. Figure 5: Architecture details of the vision transformer (ViT), which consists of multi-head self-attention layers, feed-forward networks, and residual connections for enhanced feature representation. map with a spatial resolution of 14 × 14 pixels is first partitioned into 196 non￾overlapping patches by a positional encoding layer. This creates a sequence of 196 patch embeddings, each represented as a 768-dimension… view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of our fusion model with its three component networks based on the 80:20 train-test split described in Section 2 when all the four tools are used for deterministic prediction. The fusion model achieves the best performance with a TSS score of 0.757. fusion model correctly predicts 14 of them (TN = 14) and incorrectly predicts 2 of them to be positive, although the two are actually negative (FP =… view at source ↗
Figure 7
Figure 7. Figure 7: The confusion matrix obtained by our fusion model used for deterministic prediction based on the 80:20 train-test split described in Section 2 [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of the four data types (LASCO C2, AIA 193 ˚A, AIA 211 ˚A, HMI) and their combinations, each of which is used in turn as input to our fusion model, for deterministic prediction based on the 80:20 train-test split described in Section 2. The model achieves the best performance when all the four data types are used together. importance scores to input features. The resulting attribution maps highli… view at source ↗
Figure 9
Figure 9. Figure 9: AIA 193 ˚A image (top), AIA 211 ˚A image (bottom), and their corresponding IG attribution maps for a correctly predicted (true positive) CME event that occurred on 14 March 2015 UT. The left column shows the AIA images and the right column shows the IG attribution maps. Each color bar indicates the normalized attribution intensity, ranging from 0 (blue, low attribution, low importance score) to 1 (red, hig… view at source ↗
Figure 10
Figure 10. Figure 10: AIA 193 ˚A image (top), AIA 211 ˚A image (bottom), and their corresponding IG attribution maps for an incorrectly predicted (false negative) CME event that occurred on 7 May 2019 UT. In each AIA image, the AR 12740, which is the source region of the CME event, is highlighted by a red box, while the coronal holes are outlined in green. Based on the IG attribution maps, our fusion model does not pay suffici… view at source ↗
Figure 11
Figure 11. Figure 11: Comparison of our fusion model with its three component networks based on the 80:20 train-test split described in Section 2 when all the four tools are used for probabilistic forecasting. The fusion model achieves the best performance with a BS score of 0.080. where ¯y = 1 N PN i=1 yi represents the mean frequency of actual geoeffective events in the test set. BSS values range from −∞ to 1, with a perfect… view at source ↗
Figure 12
Figure 12. Figure 12: Comparison of the four data types (LASCO C2, AIA 193 ˚A, AIA 211 ˚A, HMI) and their combinations, each of which is used in turn as input to our fusion model, for probabilistic forecasting based on the 80:20 train-test split described in Section 2. The model achieves the best performance when all the four data types are used together. 5. Discussion and Conclusion In this study, we proposed a new fusion mod… view at source ↗
read the original abstract

Understanding and forecasting the geoeffectiveness of a coronal mass ejection (CME) is crucial for protecting infrastructure in the near-Earth space environment and on Earth. In this study, we present a novel fusion model to forecast the geoeffectiveness of CME events. Our model combines convolutional neural networks for feature learning and a prediction network for feature fusion and event classification. The model is trained by observations from instruments including the Large Angle Spectroscopic Coronagraph (LASCO) on board the Solar and Heliospheric Observatory (SOHO) and the Atmospheric Imaging Assembly (AIA) and Helioseismic and Magnetic Imager (HMI) on board the Solar Dynamics Observatory (SDO). The trained model is then used to predict whether an Earth-reaching CME will cause a geomagnetic storm and/or the probability that the CME will cause such a storm. Experimental results based on a five-fold cross validation scheme demonstrate the good performance of our fusion model, achieving a mean true skill statistic (TSS) score of 0.703 when the model is used as a deterministic prediction tool, and a mean Brier score of 0.095 when the model is used as a probabilistic forecasting tool, where a TSS score of 1 or a Brier score of 0 indicates perfect performance. This work contributes to forecasting the causal relationship between Earth-directed CMEs and geomagnetic storms in solar-terrestrial interactions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces a fusion deep learning model combining CNNs on SOHO/LASCO, SDO/AIA and SDO/HMI observations to predict whether Earth-directed CMEs will produce geomagnetic storms. It reports a mean TSS of 0.703 for deterministic classification and a mean Brier score of 0.095 for probabilistic forecasting, both obtained from five-fold cross-validation on the observational data.

Significance. If the reported metrics are shown to be robust after supplying the missing dataset and labeling details, the work would provide a practical multi-instrument deep-learning pipeline for operational space-weather forecasting of CME geoeffectiveness.

major comments (2)
  1. [Abstract and Methods] Abstract and Methods section: the manuscript supplies no information on dataset size, number of events, class balance, CME selection criteria, the catalog or Dst threshold used to assign geoeffectiveness labels, or verification of label completeness. These omissions are load-bearing because the central claim (mean TSS 0.703, Brier score 0.095) cannot be evaluated without confirming that the binary and probabilistic labels are accurate and drawn from an unbiased sample of Earth-directed CMEs.
  2. [Methods] Cross-validation description (Methods): there is no statement on whether the five-fold CV respects temporal ordering of the solar observations. Without this, the reported scores may reflect temporal leakage rather than genuine predictive skill.
minor comments (1)
  1. [Abstract] The abstract could usefully state the time span of the observations used for training and testing.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which highlight important aspects of our work that require clarification. We address each major comment below and will revise the manuscript accordingly to strengthen the presentation of our methods and results.

read point-by-point responses

  1. Referee: [Abstract and Methods] Abstract and Methods section: the manuscript supplies no information on dataset size, number of events, class balance, CME selection criteria, the catalog or Dst threshold used to assign geoeffectiveness labels, or verification of label completeness. These omissions are load-bearing because the central claim (mean TSS 0.703, Brier score 0.095) cannot be evaluated without confirming that the binary and probabilistic labels are accurate and drawn from an unbiased sample of Earth-directed CMEs.

    Authors: We agree that these details are essential for proper evaluation of the reported metrics. The revised manuscript will include the dataset size, number of events, class balance, CME selection criteria, the specific catalog employed, the Dst threshold used for geoeffectiveness labeling, and verification steps for label completeness. revision: yes

  2. Referee: [Methods] Cross-validation description (Methods): there is no statement on whether the five-fold CV respects temporal ordering of the solar observations. Without this, the reported scores may reflect temporal leakage rather than genuine predictive skill.

    Authors: We recognize the critical need to address potential temporal leakage in cross-validation for time-ordered solar data. The revised manuscript will explicitly describe whether the five-fold CV respects temporal ordering and provide the rationale for the approach used, ensuring the metrics reflect genuine predictive skill. revision: yes

Circularity Check

0 steps flagged

No circularity: standard supervised ML training and CV evaluation on external labels

full rationale

The paper trains a CNN-based fusion model on SOHO/SDO image observations paired with external binary/probabilistic geoeffectiveness labels (geomagnetic storm occurrence), then reports TSS and Brier scores via 5-fold cross-validation on held-out events. No equations define model outputs or performance metrics in terms of themselves; the labels are not derived from the model; no self-citations, ansatzes, or uniqueness theorems are invoked to force the result; and the evaluation is a standard external benchmark. The derivation chain is therefore self-contained and does not reduce to its inputs by construction.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The claim rests on the domain assumption that image features alone suffice for the prediction task and on the implicit premise that the historical event labels are reliable ground truth.

free parameters (1)
  • neural network weights and architecture hyperparameters
    All model parameters are learned from the training images and labels.
axioms (1)
  • domain assumption Visual features extracted from LASCO, AIA, and HMI images contain sufficient information to determine whether an Earth-directed CME will produce a geomagnetic storm.
    This premise is required for the image-only CNN approach to be valid.

pith-pipeline@v0.9.1-grok · 5813 in / 1233 out tokens · 45489 ms · 2026-06-30T12:01:44.795203+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

57 extracted references · 54 canonical work pages

  1. [1]

    , Abduallah , Y

    barticle Alobaid , K.A. , Abduallah , Y. , Wang , J.T.L. , Wang , H. , Jiang , H. , Xu , Y. , Yurchyshyn , V. , Zhang , H. , Cavus , H. , Jing , J. : 2022 , Predicting CME arrival time through data integration and ensemble learning . Frontiers in Astronomy and Space Sciences 9 , 1013345 . https://doi.org/10.3389/fspas.2022.1013345 . barticle

    work page doi:10.3389/fspas.2022.1013345 2022

  2. [2]

    , Abduallah , Y

    barticle Alobaid , K.A. , Abduallah , Y. , Wang , J.T.L. , Wang , H. , Fan , S. , Li , J. , Cavus , H. , Yurchyshyn , V. : 2023 , Estimating coronal mass ejection mass and kinetic energy by fusion of multiple deep-learning models . 958 , L34 . https://doi.org/10.3847/2041-8213/ad0c4a . barticle

    work page doi:10.3847/2041-8213/ad0c4a 2023

  3. [3]

    , Wang , J.T.L

    barticle Alobaid , K.A. , Wang , J.T.L. , Wang , H. , Jing , J. , Abduallah , Y. , Wang , Z. , Farooki , H. , Cavus , H. , Yurchyshyn , V. : 2024 , Prediction of geoeffective CMEs using SOHO images and deep learning . 299 , 159 . https://doi.org/10.1007/s11207-024-02385-w . barticle

    work page doi:10.1007/s11207-024-02385-w 2024

  4. [4]

    , Hinterreiter , J

    barticle Amerstorfer , T. , Hinterreiter , J. , Reiss , M.A. , M \"o stl , C. , Davies , J.A. , Bailey , R.L. , Weiss , A.J. , Dumbovi \'c , M. , Bauer , M. , Amerstorfer , U.V. , Harrison , R.A. : 2021 , Evaluation of CME arrival prediction using ensemble modeling based on heliospheric imaging observations . Space Weather 19 , e02553 . https://doi.org/10...

    work page doi:10.1029/2020sw002553 2021

  5. [5]

    , Verbeke , C

    barticle Baratashvili , T. , Verbeke , C. , Wijsen , N. , Poedts , S. : 2022 , Improving CME evolution and arrival predictions with AMR and grid stretching in Icarus . 667 , A133 . https://doi.org/10.1051/0004-6361/202244111 . barticle

    work page doi:10.1051/0004-6361/202244111 2022

  6. [6]

    , Talpeanu , D.-C

    barticle Besliu-Ionescu , D. , Talpeanu , D.-C. , Mierla , M. , Maris Muntean , G. : 2019 , On the prediction of geoeffectiveness of CMEs during the ascending phase of SC24 using a logistic regression method . J. Atmos. Solar-Terr. Phys. 193 , 105036 . https://doi.org/10.1016/j.jastp.2019.04.017 . barticle

    work page doi:10.1016/j.jastp.2019.04.017 2019

  7. [7]

    , Higgins , P.A

    barticle Bloomfield , D.S. , Higgins , P.A. , McAteer , R.T.J. , Gallagher , P.T. : 2012 , Toward reliable benchmarking of solar flare forecasting methods . 747 , L41 . https://doi.org/10.1088/2041-8205/747/2/L41 . barticle

    work page doi:10.1088/2041-8205/747/2/l41 2012

  8. [8]

    : 1950 , Verification of forecasts expressed in terms of probability

    barticle Brier , G.W. : 1950 , Verification of forecasts expressed in terms of probability . Monthly Weather Review 78 , 1 . https://doi.org/10.1175/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2 . barticle

    work page doi:10.1175/1520-0493(1950)078 1950

  9. [9]

    , Howard , R.A

    barticle Brueckner , G.E. , Howard , R.A. , Koomen , M.J. , Korendyke , C.M. , Michels , D.J. , Moses , J.D. , Socker , D.G. , Dere , K.P. , Lamy , P.L. , Llebaria , A. , Bout , M.V. , Schwenn , R. , Simnett , G.M. , Bedford , D.K. , Eyles , C.J. : 1995 , The Large Angle Spectroscopic Coronagraph (LASCO) . 162 , 357 . https://doi.org/10.1007/BF00733434 . barticle

    work page doi:10.1007/bf00733434 1995

  10. [10]

    , Sachdeva , N

    barticle Chen , H. , Sachdeva , N. , Huang , Z. , Holst , B. , Manchester , W. , Jivani , A. , Zou , S. , Chen , Y. , Huan , X. , Toth , G. : 2025 , Decent estimate of CME arrival time from a data-assimilated ensemble in the Alfv \'e n Wave Solar atmosphere Model (DECADE-AWSoM) . Space Weather 23 , 2024SW004165 . https://doi.org/10.1029/2024SW004165 . barticle

    work page doi:10.1029/2024sw004165 2025

  11. [11]

    , Liu , J

    barticle Chierichini , S. , Liu , J. , Kors \'o s , M.B. , Del Moro , D. , Erd \'e lyi , R. : 2024 , CME arrival modeling with machine learning . 963 , 121 . https://doi.org/10.3847/1538-4357/ad1cee . barticle

    work page doi:10.3847/1538-4357/ad1cee 2024

  12. [12]

    ImageNet: A Large- Scale Hierarchical Image Database

    bchapter Deng , J. , Dong , W. , Socher , R. , Li , L.-J. , Li , K. , Fei-Fei , L. : 2009 , ImageNet: A large-scale hierarchical image database . In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 248 . https://doi.org/10.1109/CVPR.2009.5206848 . bchapter

    work page doi:10.1109/cvpr.2009.5206848 2009

  13. [13]

    , Veronig , A.M

    barticle Dissauer , K. , Veronig , A.M. , Temmer , M. , Podladchikova , T. , Vanninathan , K. : 2018 , Statistics of coronal dimmings associated with coronal mass ejections. I. Characteristic dimming properties and flare association . 863 , 169 . https://doi.org/10.3847/1538-4357/aad3c6 . barticle

    work page doi:10.3847/1538-4357/aad3c6 2018

  14. [14]

    , Veronig , A.M

    barticle Dissauer , K. , Veronig , A.M. , Temmer , M. , Podladchikova , T. : 2019 , Statistics of coronal dimmings associated with coronal mass ejections. II. Relationship between coronal dimmings and their associated CMEs . 874 , 123 . https://doi.org/10.3847/1538-4357/ab0962 . barticle

    work page doi:10.3847/1538-4357/ab0962 2019

  15. [15]

    , Fleck , B

    barticle Domingo , V. , Fleck , B. , Poland , A.I. : 1995 , The SOHO mission: An overview . 162 , 1 . https://doi.org/10.1007/BF00733425 . barticle

    work page doi:10.1007/bf00733425 1995

  16. [16]

    , Beyer , L

    bchapter Dosovitskiy , A. , Beyer , L. , Kolesnikov , A. , Weissenborn , D. , Zhai , X. , Unterthiner , T. , Dehghani , M. , Minderer , M. , Heigold , G. , Gelly , S. , Uszkoreit , J. , Houlsby , N. : 2021 , An image is worth 16x16 words: Transformers for image recognition at scale . In: Proceedings of the 9th International Conference on Learning Represen...

    2021

  17. [17]

    , C alogovi \'c , J

    barticle Dumbovi \'c , M. , C alogovi \'c , J. , Martini \'c , K. , Vr s nak , B. , Sudar , D. , Temmer , M. , Veronig , A. : 2021 , Drag-based model (DBM) tools for forecast of coronal mass ejection arrival time and speed . Front. Astron. Space Sci. 8 , 58 . https://doi.org/10.3389/fspas.2021.639986 . barticle

    work page doi:10.3389/fspas.2021.639986 2021

  18. [18]

    , Moore , R.L

    barticle Falconer , D.A. , Moore , R.L. , Gary , G.A. : 2006 , Magnetic causes of solar coronal mass ejections: Dominance of the free magnetic energy over the magnetic twist alone . 644 , 1258 . https://doi.org/10.1086/503699 . barticle

    work page doi:10.1086/503699 2006

  19. [19]

    , Zheng , Y

    barticle Fu , H. , Zheng , Y. , Ye , Y. , Feng , X. , Liu , C. , Ma , H. : 2021 , Joint geoeffectiveness and arrival time prediction of CMEs by a unified deep learning framework . Remote Sens. 13 , 1738 . https://doi.org/10.3390/rs13091738 . barticle

    work page doi:10.3390/rs13091738 2021

  20. [20]

    , Joselyn , J.A

    barticle Gonzalez , W.D. , Joselyn , J.A. , Kamide , Y. , Kroehl , H.W. , Rostoker , G. , Tsurutani , B.T. , Vasyliunas , V.M. : 1994 , What is a geomagnetic storm? 99 , 5771 . https://doi.org/10.1029/93JA02867 . barticle

    work page doi:10.1029/93ja02867 1994

  21. [21]

    : 2009 , Halo coronal mass ejections and geomagnetic storms

    barticle Gopalswamy , N. : 2009 , Halo coronal mass ejections and geomagnetic storms . Earth, Planets and Space 61 , 595 . https://doi.org/10.1186/BF03352930 . barticle

    work page doi:10.1186/bf03352930 2009

  22. [22]

    , M \"a kel \"a , P

    barticle Gopalswamy , N. , M \"a kel \"a , P. , Xie , H. , Akiyama , S. , Yashiro , S. : 2009 , CME interactions with coronal holes and their interplanetary consequences . Journal of Geophysical Research (Space Physics) 114 , A00A22 . https://doi.org/10.1029/2008JA013686 . barticle

    work page doi:10.1029/2008ja013686 2009

  23. [23]

    , Piana , M

    barticle Guastavino , S. , Piana , M. , Massone , A.M. , Schwartz , R. , Benvenuto , F. : 2019 , Desaturating SDO/AIA observations of solar flaring storms . 882 , 109 . https://doi.org/10.3847/1538-4357/ab35d8 . barticle

    work page doi:10.3847/1538-4357/ab35d8 2019

  24. [24]

    , Candiani , V

    barticle Guastavino , S. , Candiani , V. , Bemporad , A. , Marchetti , F. , Benvenuto , F. , Massone , A.M. , Mancuso , S. , Susino , R. , Telloni , D. , Fineschi , S. , Piana , M. : 2023 , Physics-driven machine learning for the prediction of coronal mass ejections' travel times . 954 , 151 . https://doi.org/10.3847/1538-4357/ace62d . barticle

    work page doi:10.3847/1538-4357/ace62d 2023

  25. [25]

    , Bahamazava , K

    barticle Guastavino , S. , Bahamazava , K. , Perracchione , E. , Camattari , F. , Audone , G. , Telloni , D. , Susino , R. , Nicolini , G. , Fineschi , S. , Piana , M. , Massone , A.M. : 2024 , Forecasting geoffective events from solar wind data and evaluating the most predictive features through machine learning approaches . 971 , 94 . https://doi.org/10...

    work page doi:10.3847/1538-4357/ad5b57 2024

  26. [26]

    Deep residual learning for image recognition

    bchapter He , K. , Zhang , X. , Ren , S. , Sun , J. : 2016 , Deep residual learning for image recognition . In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 770 . https://doi.org/10.1109/CVPR.2016.90 . bchapter

    work page doi:10.1109/cvpr.2016.90 2016

  27. [27]

    , Camporeale , E

    barticle Hu , A. , Camporeale , E. , Swiger , B. : 2023 , Multi‐hour‐ahead Dst index prediction using multi‐fidelity boosted neural networks . Space Weather 21 . https://doi.org/10.1029/2022sw003286 . barticle

    work page doi:10.1029/2022sw003286 2023

  28. [28]

    , Cheung , M

    barticle Hurlburt , N. , Cheung , M. , Schrijver , C. , Chang , L. , Freeland , S. , Green , S. , Heck , C. , Jaffey , A. , Kobashi , A. , Schiff , D. , Serafin , J. , Seguin , R. , Slater , G. , Somani , A. , Timmons , R. : 2012 , Heliophysics Event Knowledgebase for the Solar Dynamics Observatory (SDO) and beyond . 275 , 67 . https://doi.org/10.1007/s11...

    work page doi:10.1007/s11207-010-9624-2 2012

  29. [29]

    , Shugay , Y.S

    barticle Kaportseva , K.B. , Shugay , Y.S. : 2021 , Use of the DBM model to the predict of arrival of coronal mass ejections to the Earth . Cosmic Res. 59 , 268 . https://doi.org/10.1134/S001095252104002X . barticle

    work page doi:10.1134/s001095252104002x 2021

  30. [30]

    The Atmospheric Imaging Assembly (AIA) on the Solar Dynamics Observatory (SDO).Sol

    barticle Lemen , J.R. , Title , A.M. , Akin , D.J. , Boerner , P.F. , Chou , C. , Drake , J.F. , Duncan , D.W. , Edwards , C.G. , Friedlaender , F.M. , Heyman , G.F. , Hurlburt , N.E. , Katz , N.L. , Kushner , G.D. , Levay , M. , Lindgren , R.W. , Mathur , D.P. , McFeaters , E.L. , Mitchell , S. , Rehse , R.A. , Schrijver , C.J. , Springer , L.A. , Stern ...

    work page doi:10.1007/s11207-011-9776-8 2012

  31. [31]

    , Ye , Y

    barticle Liu , J. , Ye , Y. , Shen , C. , Wang , Y. , Erd \'e lyi , R. : 2018 , A new tool for CME arrival time prediction using machine learning algorithms: CAT-PUMA . 855 , 109 . https://doi.org/10.3847/1538-4357/aaae69 . barticle

    work page doi:10.3847/1538-4357/aaae69 2018

  32. [32]

    , Shen , C

    barticle Liu , J. , Shen , C. , Wang , Y. , Xu , M. , Chi , Y. , Zhong , Z. , Mao , D. , Zhang , Z. , Wang , C. , Liu , J. , Wang , Y. : 2024 , Forecasting the Dst index with temporal convolutional network and integrated gradients . 299 , 98 . https://doi.org/10.1007/s11207-024-02340-9 . barticle

    work page doi:10.1007/s11207-024-02340-9 2024

  33. [33]

    , Webb , D.F

    barticle Liu , Y. , Webb , D.F. , Zhao , X.P. : 2006 , Magnetic structures of solar active regions, full-halo coronal mass ejections, and geomagnetic storms . 646 , 1335 . https://doi.org/10.1086/505036 . barticle

    work page doi:10.1086/505036 2006

  34. [34]

    : 1980 , What is a geomagnetic index? In: Derivation, Meaning, and Use of Geomagnetic Indices , American Geophysical Union

    bchapter Mayaud , P.N. : 1980 , What is a geomagnetic index? In: Derivation, Meaning, and Use of Geomagnetic Indices , American Geophysical Union . Chap. 2 . https://doi.org/10.1002/9781118663837.ch2 . bchapter

    work page doi:10.1002/9781118663837.ch2 1980

  35. [35]

    , Yang , Q

    barticle Pan , S.J. , Yang , Q. : 2010 , A survey on transfer learning . IEEE Trans. Knowl. Data Eng. 22 , 1345 . https://doi.org/10.1109/TKDE.2009.191 . barticle

    work page doi:10.1109/tkde.2009.191 2010

  36. [36]

    , Micha ek , G

    barticle Pattnaik , A. , Micha ek , G. , Ravishankar , A. : 2025 , Estimation of transit time of CMEs from the Sun to Earth . 982 , 35 . https://doi.org/10.3847/1538-4357/adb84b . barticle

    work page doi:10.3847/1538-4357/adb84b 2025

  37. [37]

    , Thompson , B.J

    barticle Pesnell , W.D. , Thompson , B.J. , Chamberlin , P.C. : 2012 , The Solar Dynamics Observatory (SDO) . 275 , 3 . https://doi.org/10.1007/s11207-011-9841-3 . barticle

    work page doi:10.1007/s11207-011-9841-3 2012

  38. [38]

    , Paraschiv , A.R

    barticle Pricopi , A.-C. , Paraschiv , A.R. , Besliu-Ionescu , D. , Marginean , A.-N. : 2022 , Predicting the geoeffectiveness of CMEs using machine learning . 934 , 176 . https://doi.org/10.3847/1538-4357/ac7962 . barticle

    work page doi:10.3847/1538-4357/ac7962 2022

  39. [39]

    , Cane , H.V

    barticle Richardson , I.G. , Cane , H.V. : 2010 , Near- Earth interplanetary coronal mass ejections during solar cycle 23 (1996 - 2009): Catalog and summary of properties . 264 , 189 . https://doi.org/10.1007/s11207-010-9568-6 . barticle

    work page doi:10.1007/s11207-010-9568-6 2010

  40. [40]

    , Schou , J

    barticle Scherrer , P.H. , Schou , J. , Bush , R.I. , Kosovichev , A.G. , Bogart , R.S. , Hoeksema , J.T. , Liu , Y. , Duvall , T.L. , Zhao , J. , Title , A.M. , Schrijver , C.J. , Tarbell , T.D. , Tomczyk , S. : 2012 , The Helioseismic and Magnetic Imager (HMI) investigation for the Solar Dynamics Observatory (SDO) . 275 , 207 . https://doi.org/10.1007/s...

    work page doi:10.1007/s11207-011-9834-2 2012

  41. [41]

    , Torre , G

    barticle Schwartz , R.A. , Torre , G. , Piana , M. : 2014 , Systematic de-saturation of images from the Atmospheric Imaging Assembly in the Solar Dynamics Observatory . 793 , L23 . https://doi.org/10.1088/2041-8205/793/2/L23 . barticle

    work page doi:10.1088/2041-8205/793/2/l23 2014

  42. [42]

    , Torre , G

    barticle Schwartz , R.A. , Torre , G. , Massone , A.M. , Piana , M. : 2015 , DESAT: A Solar SoftWare tool for image de-saturation in the Atmospheric Image Assembly onboard the Solar Dynamics Observatory . Astronomy and Computing 13 , 117 . https://doi.org/10.1016/j.ascom.2015.10.006 . barticle

    work page doi:10.1016/j.ascom.2015.10.006 2015

  43. [43]

    , Vr s nak , B

    barticle Sudar , D. , Vr s nak , B. , Dumbovi \'c , M. : 2016 , Predicting coronal mass ejections transit times to Earth with neural network . 456 , 1542 . https://doi.org/10.1093/mnras/stv2782 . barticle

    work page doi:10.1093/mnras/stv2782 2016

  44. [44]

    , Taly , A

    bchapter Sundararajan , M. , Taly , A. , Yan , Q. : 2017 , Axiomatic attribution for deep networks . In: Proceedings of the 34th International Conference on Machine Learning , 3319 . bchapter

    2017

  45. [45]

    , Christe , S

    barticle SunPy Community , Mumford , S.J. , Christe , S. , P \'e rez-Su \'a rez , D. , Ireland , J. , Shih , A.Y. , Inglis , A.R. , Liedtke , S. , Hewett , R.J. , Mayer , F. , Hughitt , K. , Freij , N. , Meszaros , T. , Bennett , S.M. , Malocha , M. , Evans , J. , Agrawal , A. , Leonard , A.J. , Robitaille , T.P. , Mampaey , B. , Campos-Rozo , J.I. , Kirk...

    work page doi:10.1088/1749-4699/8/1/014009 2015

  46. [46]

    , Le , Q.V

    bchapter Tan , M. , Le , Q.V. : 2019 , EfficientNet: Rethinking model scaling for convolutional neural networks . In: Proceedings of the 36th International Conference on Machine Learning 97 , 6105 . bchapter

    2019

  47. [47]

    : 2022 , Statistical methods applied to space weather science

    barticle Telloni , D. : 2022 , Statistical methods applied to space weather science . Front. Astron. Space Sci. 9 , 865880 . https://doi.org/10.3389/fspas.2022.865880 . barticle

    work page doi:10.3389/fspas.2022.865880 2022

  48. [48]

    , Schiavo , M.L

    barticle Telloni , D. , Schiavo , M.L. , Magli , E. , Fineschi , S. , Guastavino , S. , Nicolini , G. , Susino , R. , Giordano , S. , Amadori , F. , Candiani , V. , Massone , A.M. , Piana , M. : 2023 , Prediction capability of geomagnetic events from solar wind data using neural networks . 952 , 111 . https://doi.org/10.3847/1538-4357/acdeea . barticle

    work page doi:10.3847/1538-4357/acdeea 2023

  49. [49]

    , Wang , H

    barticle Toriumi , S. , Wang , H. : 2019 , Flare-productive active regions . Living Reviews in Solar Physics 16 , 3 . https://doi.org/10.1007/s41116-019-0019-7 . barticle

    work page doi:10.1007/s41116-019-0019-7 2019

  50. [50]

    , Patsourakos , S

    barticle Vourlidas , A. , Patsourakos , S. , Savani , N.P. : 2019 , Predicting the geoeffective properties of coronal mass ejections: current status, open issues and path forward . Philosophical Transactions of the Royal Society A 377 , 20180096 . https://doi.org/10.1098/rsta.2018.0096 . barticle

    work page doi:10.1098/rsta.2018.0096 2019

  51. [51]

    , Liu , J

    barticle Wang , Y. , Liu , J. , Jiang , Y. , Erd \'e lyi , R. : 2019 , CME arrival time prediction using convolutional neural network . 881 , 15 . https://doi.org/10.3847/1538-4357/ab2b3e . barticle

    work page doi:10.3847/1538-4357/ab2b3e 2019

  52. [52]

    , Showalter , K.M

    barticle Wanliss , J.A. , Showalter , K.M. : 2006 , High-resolution global storm index: Dst versus SYM-H . J. Geophys. Res. (Space Physics) 111 , A02202 . https://doi.org/10.1029/2005JA011034 . barticle

    work page doi:10.1029/2005ja011034 2006

  53. [53]

    : 2010 , Forecast verification

    bchapter Wilks , D.S. : 2010 , Forecast verification . In: Statistical Methods in the Atmospheric Sciences , 3 rd edn., Elsevier . Chap. 8 . https://doi.org/10.1016/B978-0-12-385022-5.00008-7 . bchapter

    work page doi:10.1016/b978-0-12-385022-5.00008-7 2010

  54. [54]

    , Li , H

    barticle Ye , D. , Li , H. , Guo , L. , Jiang , X. : 2025 , An improved halo coronal mass ejection geoeffectiveness prediction model using multiple coronal mass ejection features based on the DC-PCA-KNN method . 978 , 66 . https://doi.org/10.3847/1538-4357/ad98f0 . barticle

    work page doi:10.3847/1538-4357/ad98f0 2025

  55. [55]

    , Liu , J

    barticle Ye , Y. , Liu , J. , Hao , Y. , Cui , J. : 2024 , Evaluating the geoeffectiveness of interplanetary coronal mass ejections: Insights from a support vector machine approach with SHAP value analysis . 972 , 52 . https://doi.org/10.3847/1538-4357/ad61d7 . barticle

    work page doi:10.3847/1538-4357/ad61d7 2024

  56. [56]

    , Xu , L

    barticle Yu , X. , Xu , L. , Ren , Z. , Zhao , D. , Sun , W. : 2022 , Image desaturation for SDO/AIA using mixed convolution network . Research in Astronomy and Astrophysics 22 , 065009 . https://doi.org/10.1088/1674-4527/ac69b7 . barticle

    work page doi:10.1088/1674-4527/ac69b7 2022

  57. [57]

    , Jing , J

    barticle Zhang , H. , Jing , J. , Wang , J.T.L. , Wang , H. , Abduallah , Y. , Xu , Y. , Alobaid , K.A. , Farooki , H. , Yurchyshyn , V. : 2025 , Prediction of halo coronal mass ejections using SDO/HMI vector magnetic data products and a transformer model . 981 , 37 . https://doi.org/10.3847/1538-4357/adafa0 . barticle

    work page doi:10.3847/1538-4357/adafa0 2025