We’ve got another new paper from the UW Massive Stars research group! Led by UW grad student Trevor Dorn-Wallenstein, this paper started with a large database of optical and IR photometry (including data from Gaia, WISE, and 2MASS) and then tested several different machine learning methods for trying to spectroscopically classify these stars based on their photometry and light curves. The results showed that a Support Vector Machine is capable of coarsely classifying massive stars into “hot”, “cool” and “emission line categories”, and has a 76% success rate of identifying emission line stars without any observations of the stars’ spectra or emission lines themselves! The paper classifies ~2500 stars that don’t have any existing labels, including fourteen new candidate emission line objects. These results have exciting prospects for the coming era of massive star observations with the Webb and Roman space telescopes, opening up the possibility of classifying them based on photometry alone.
The paper is in press with the Astrophysical Journal; for now you can check it out on arXiv!
The post New paper: classifying massive stars with machine learning! first appeared on E. M. Levesque.
