Hete subluminieuze sterren: Tijdreeksanalyse en pulsatiemodus-identificatie

This project aims to understand the population and physics of variable stars from multi-dimensional data assembled with BlackGEM on the one hand, and available in public databases on the other hand, by using machine learning techniques. A secondary goal is to understand how ground-based and space-based data can properly support each other in the understanding of stellar variability and the underlying stellar structure and evolution. The challenge is to use multi-dimensional, heterogeneous, heteroskedastic data to constrain/identify/interpret variability in stars, covering many orders of apparent magnitude to understand the variable sky. We will investigate complementarity of data sources, overlap in data sources, gaps in data sources, limiting magnitudes/fluxes, etc. Propagation of errors in statistical models between data sets will be assessed, as pathways towards optimal combination of various input data sets and their uncertainties to interpret stellar variability at large scale.