Machine learning classification of RR Lyrae stars observed by TESS

In this study we present first results of the investigation of different ML-classifiers and their ability to distinguish between different subclasses of RR Lyrae stars (RRab and RRc) from the morphology of TESS-lightcurves. From TESS full-frame images we extracted over 3000 lightcurves of stars listed as RRab and RRc in the General Catalogue of Variable Stars (GCVS). The extracted lightcurves were preprocessed and analyzed with a dedicated python-package. For more than 100 RR Lyrae stars, we determined pulsation periods that are not yet listed in the GCVS and verified the given periods of the remaining ones. On the extracted lightcurves, we test and compare the performance of three unsupervised clustering algorithms (DBSCAN, HDBSCAN, KMeans) in combination with different projection techniques (t-SNE, UMAP) against the classes provided in the GCVS. Independent observations from other surveys such as GAIA will further improve the accuracy of this automated classification procedure, which in the future we plan to apply to different pulsator-classes.