Deep learning for early warning signals of tipping points : supplementary data

This data accompanies the publication by Bury et al. “Deep learning for early warning signals of tipping points” published in PNAS and the Github repository https://github.com/ThomasMBury/deep-early-warnings-pnas. It contains the model time series data that are used to train the deep learning algorithm. The directory ts_500 contains 500k time series used to train the 500-classifier. The directory ts_1500 contains 200k time series used to train the 1500-classifier. Both directories contain files <em>labels.csv </em>and <em>groups.csv </em>which provide numbers corresponding to the labels (Fold, Hopf, Branch, Null) and groups (Training, Validation, Test) for each time series respectively.