Neural Network identification of halo white dwarfs

Abstract

The white dwarf luminosity function has proven to be an excellent tool to study some properties of the galactic disk such as its age and the past history of the local star formation rate. The existence of an observational luminosity function for halo white dwarfs could provide valuable information about its age, the time that the star formation rate lasted, and could also constrain the shape of the allowed Initial Mass Functions (IMF). However, the main problem is the scarce number of white dwarfs already identified as halo stars. In this Letter we show how an artificial intelligence algorithm can be succesfully used to classify the population of spectroscopically identified white dwarfs allowing us to identify several potential halo white dwarfs and to improve the significance of its luminosity function.