The group has developed a fast and precise tool to determine stellar parameters through the use of machine learning. They use random forest regression to measure evolutionary, structural and chemical stellar attributes from observable features of stars, such as asteroseismic observables. The random forest provides information about so-called feature importance, i.e. it indicates what observables are used most often in decisions made by the trees in the random forest. This is essential feedback to know where to focus observational efforts on.
Asteroseismology and beyond
The power of asteroseismology reaches farther than the structure and properties of stars. It also impacts significantly on our knowledge of extra-solar planets, the Milky Way and as mentioned before the expansion of the universe. The determination of the planetary properties (mass, radius and age) critically relies on the knowledge of the same properties of the host star. As the ebook explains, it was thanks to asteroseismology that it was possible to detect Earth-mass planets in the habitable zone.
To utilise the full extent of the power of asteroseismology it is essential to fully understand all the oscillation features that are present in the data and understand their physical origin. The work of the Group is still in its infancy and will require many in-depth asteroseismic studies as the field matures.