Understanding nonlinearities with Machine-Learning techniques
Abstract
Ultra-precise photometric data gathered by recent space missions allow to improve our knowledge of stars significantly but challenges to find models that fit the observations with the state-of-the-art theory are raised. In addition, the huge amount of data requires the application of innovative data analysis techniques to exploit the information. Machine learning could help us both to deal with the large amount of data and to explore new ways to process the lightcurves of pulsating stars. Of special interest for studies of intermediate mass pulsating stars is the decades-old unsolved non-linear phenomenon. The aim of this project is to use machine learning tools in order to characterize a sample of pulsating stars through clustering analysis to shed some light into the nature of the nonlinear phenomenon. Specifically, we focus here on the classification of Delta Scuti stars as High Amplitude (HADS) and Low Amplitude (LADS) through the analysis of the nonlinear parameters characterizing the lightcurves (e.g. harmonics and combinations). The input parameters are extracted from the lightcurves of a sample of stars observed by TESS space satellite. The application of ML for understanding nonlinearities have a great potential use for the Complementary Science Program of PLATO.
- Publication:
-
PLATO Stellar Science Conference 2023
- Pub Date:
- July 2023
- DOI:
- 10.5281/zenodo.8173480
- Bibcode:
- 2023plat.confE...3P
- Keywords:
-
- Delta Scuti;
- Nonlinearities;
- Stellar Pulsation;
- Data Analysis;
- Machine Learning;
- Zenodo community pssc23