Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PRESOL: a web-based computational setting for feature-based flare forecasting

Chiara Curletto, Paolo Massa, Valeria Tagliafico, Cristina Campi, Federico Benvenuto, Michele Piana, Andrea Tacchino

TL;DR

The paper tackles solar flare forecasting by advocating a feature-based ML approach and presents PRESOL, a web-based, zero-footprint platform that automates SHARP feature extraction, data annotation via GOES/HEK, dataset splitting, multiple ML algorithms, feature ranking with Recursive Feature Elimination, and comprehensive performance metrics. It demonstrates an end-to-end, modular workflow with a user-friendly interface and robust evaluation across repeated train–test splits, highlighting how feature correlations influence rankings and how removing them stabilizes predictions. The work provides a path toward interpretable, reproducible space-weather forecasting with potential operational utility, and emphasizes extensibility to incorporate new algorithms and data sources. Overall, PRESOL offers a transparent, scalable framework for feature-driven flare prediction that complements more opaque deep-learning approaches and supports practical decision-making in space weather.

Abstract

Solar flares are the most explosive phenomena in the solar system and the main trigger of the events' chain that starts from Coronal Mass Ejections and leads to geomagnetic storms with possible impacts on the infrastructures at Earth. Data-driven solar flare forecasting relies on either deep learning approaches, which are operationally promising but with a low explainability degree, or machine learning algorithms, which can provide information on the physical descriptors that mostly impact the prediction. This paper describes a web-based technological platform for the execution of a computational pipeline of feature-based machine learning methods that provide predictions of the flare occurrence, feature ranking information, and assessment of the prediction performances.

PRESOL: a web-based computational setting for feature-based flare forecasting

TL;DR

The paper tackles solar flare forecasting by advocating a feature-based ML approach and presents PRESOL, a web-based, zero-footprint platform that automates SHARP feature extraction, data annotation via GOES/HEK, dataset splitting, multiple ML algorithms, feature ranking with Recursive Feature Elimination, and comprehensive performance metrics. It demonstrates an end-to-end, modular workflow with a user-friendly interface and robust evaluation across repeated train–test splits, highlighting how feature correlations influence rankings and how removing them stabilizes predictions. The work provides a path toward interpretable, reproducible space-weather forecasting with potential operational utility, and emphasizes extensibility to incorporate new algorithms and data sources. Overall, PRESOL offers a transparent, scalable framework for feature-driven flare prediction that complements more opaque deep-learning approaches and supports practical decision-making in space weather.

Abstract

Solar flares are the most explosive phenomena in the solar system and the main trigger of the events' chain that starts from Coronal Mass Ejections and leads to geomagnetic storms with possible impacts on the infrastructures at Earth. Data-driven solar flare forecasting relies on either deep learning approaches, which are operationally promising but with a low explainability degree, or machine learning algorithms, which can provide information on the physical descriptors that mostly impact the prediction. This paper describes a web-based technological platform for the execution of a computational pipeline of feature-based machine learning methods that provide predictions of the flare occurrence, feature ranking information, and assessment of the prediction performances.

Paper Structure

This paper contains 9 sections, 1 equation, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. The source data: SHARP features
  3. Dataset preparation
  4. Annotation
  5. Splitting
  6. Machine learning algorithms and their validation
  7. Technological framework
  8. Results
  9. Conclusions

Figures (5)

  • Figure 1: Snapshot of the PRESOL interface with some miners on the left.
  • Figure 2: The PRESOL interface previewing all available datasets that can be used for training.
  • Figure 3: Feature correlation on the whole dataset
  • Figure 4: TSS values on the test set using different thresholding optimization.
  • Figure 5: TSS values of the model, without the highly correlated features, on the test set using different thresholding optimization.