zigongxu / espd2024

Machine Learning lecture at the 1st European solar physics division (ESPD) summerschool: Energisation and heating in the solar plasma 29 April - 3 May 2024 Dubrovnik, Croatia

ESPD2024

A Machine Learning tutorial on how to build your Flare and CME predictor at the 1st European Solar Physics Division (ESPD) summer school: Energisation and heating in the solar plasma 29 April - 3 May 2024 Dubrovnik, Croatia. Google slides with the content available in here.

Run locally:

To run this notebook locally, we suggest that you create a new conda environement using the environment.yml file in this repository. This can be done by first cloning this repository i.e. git clone https://github.com/cdiazbas/ESPD2024 and then creating a new environment as such from the ESPD2024 folder:

$ conda env create -f environment.yml
$ conda activate espd_machinelearning
$ jupyter lab

The jupyter lab will open our repository with all the notebooks in your browser.

If you already have a conda environment, simply install the packages using pip install -r requirements.txt.

If you need to remove the environment, simply run the following command: conda env remove -n espd_machinelearning.

Run using Google Colab:

You can run the notebooks in this repository using Google Colab by clicking on the following links:

• Flare Prediction
• CME Prediction
• CME arrival time

Remember that changes in the notebooks will not be saved, unless you copy the notebooks to your Google Drive.

Run using BinderHub:

If you want you can run this notebook using Binder by clicking on this link: . It may take a few minutes to load up when you click on it for the first time