Applying Machine Learning Approaches to White Dwarf Data

The quantity of data on white dwarf stars has only recently reached a volume that necessitates the use of machine learning to analyse it, and as such little work has been done in this area. The strong gravitational fields around these dense objects leads heavier elements to very quickly (sometimes within days) sink to the centre,

leaving a pristine atmosphere of hydrogen and/or helium. However, some white dwarfs show rock-forming elements such as calcium or magnesium in their atmospheres -- evidence of the recent accretion of planetary/cometary/asteroidal material. Such "autopsies" are the only reliable way of estimating the interior

composition of planetary material outside our solar system: for a living planet one can at best only find the overall density, and estimates of the composition are then very difficult. This "post-mortem" analysis in white dwarf atmospheres is therefore a crucial tool in understanding the foundation, formation, and fate of planetary systems.