https://github.com/byte7/solar-flares-prediction-rhessi-mission
Solar flare prediction RHESSI Mission
https://github.com/byte7/solar-flares-prediction-rhessi-mission
solar-flare space-science
Last synced: about 2 months ago
JSON representation
Solar flare prediction RHESSI Mission
- Host: GitHub
- URL: https://github.com/byte7/solar-flares-prediction-rhessi-mission
- Owner: Byte7
- License: mit
- Created: 2018-02-23T17:20:56.000Z (about 7 years ago)
- Default Branch: master
- Last Pushed: 2018-04-23T17:11:21.000Z (about 7 years ago)
- Last Synced: 2025-01-31T01:16:19.400Z (4 months ago)
- Topics: solar-flare, space-science
- Language: Jupyter Notebook
- Homepage: https://byte7.github.io/blog/Predicting-SolarFlares/
- Size: 5.82 MB
- Stars: 3
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Solar-Flares-Prediction-RHESSI-Mission
## The initial steps taken in the preprocessing are as follows -
1. The dataset was taken from NASA RHESSI Data [repository](https://hesperia.gsfc.nasa.gov/rhessi_data_search/rhessi_data_search_vme.html) accessed through a web form.
2. The initial format of files was FITS. This was converted to CSV by using tools provided [here](https://fits.gsfc.nasa.gov/fits_viewer.html).
3. The many csv files generated from the FITS files were concatenated to make the current dataset in the [repository](https://github.com/Byte7/Solar-Flares-RHESSI-Mission/tree/master/data). [This](https://github.com/Byte7/Solar-Flares-RHESSI-Mission/tree/master/Join_script) script was used for the same.
## Results
Our Top 3 models' results for prediction of a energy range of a solar flare based on attributes of the flare are -
1. Gradient Boosting Classifier - 87 % accuracy
2. Random Forests Classifier - 86 % accuracy
3. Decision Tree Calssifier - 82 % accuracy
### A detailed blog post is available [here](https://byte7.github.io/blog/Predicting-SolarFlares/) explaining how I went around the data and prediction of flare's energy.