Neural network model that predicts electron flux in the plasma sheet from solar wind inputs.
This repository contains the source code and example usage for the model
that is described in the paper:
Swiger et al. (2022). Energetic Electron Flux Predictions in the
near-Earth Plasma Sheet from Solar Wind Driving. Space Weather, (under
review).
This repository is also available on zenodo.org.
All of the code is written in python; the conda package configuration file is
swpsnn.yml.
The Jupyter Notebook model_usage_example.ipynb walks through an example of
how to go from having zero data to having a trained, neural network model.
It shows how to create the model feature arrays (inputs) and model target arrays
(outputs) from OMNI, FISM-2, and THEMIS data. Then it uses the feature and
target arrays to train a neural network.
Note that the model that is trained in model_usage_example.ipynb is
only an example.
The full, trained model that is described and analyzed in the
Swiger et al., 2022 paper is located at
Model/swpsnn_v1.2.2.h5. To open and use it, follow the same steps that
are shown in Section 4 of model_usage_example.ipynb.
The model expects the input array to be in the same format as that created in
Section 2.4 of model_usage_example.ipynb.