Data of the ozone imputation software (https://gitlab.jsc.fz-juelich.de/esde/machine-learning/ozone-imputation),
see README of this repository for usage and information.
Structure:
- 'aq-bench' contains the AQ-Bench dataset, which can be used to run graph baseline experiments
- 'models' contains the predictions of simple models and correct and smooth.
- 'uba_graph_dataset' contains the pytorch geometric dataset of this study
- 'uba_graph_preproc' contains all .csv files which are used to create the pytorch geometric dataset
- 'uba_graph_raw' contains all raw data that were preprocessed to create the pytorch geometric dataset
This file structure should be maintained when using the software. Just set the variable resources_dir to the parent directory in the file source/settings.py.