The helios module of the PlasmaML repository contains utilities for
training tensorflow based models on solar data sets.
This repo is also the home of the reference implementation of the Dynamic Time Lag Regression framework and its application on the solar wind prediction task.
For more information on the helios API refer to the scaladoc.
Run the DTLR cross validation experiment with
./helios/scripts/csss_cv_runner.sh <experiment id>
After the DTLR results are stored in the folder <experiment id>, run the fixed lag baseline model.
./helios/scripts/csss_cv_bs_runner.sh <top dir> <experiment id>