Stellar Spectra Models Classification and Parameter Estimation Using Machine Learning Algorithms

Abstract

The growth of sky surveys and the large amount of stellar spectra in the current databases, has generated the necessity of developing new methods to estimate atmospheric parameters, a fundamental task on stellar research. In this work we present a comparison of different machine learning algorithms, using for the classification of stellar synthetic spectra and the estimation of fundamental stellar parameters included T_eff(K), log(L/Lo), log g, M/Mo, and Vrot. For both tasks, we established a group of supervised learning models, and propose a database of measures with the same structure to train the algorithms. This data includes equivalent-width types measurements over noisy synthetic spectra in order to replicate the natural noise on a real observed spectrum. Different levels of signal to noise ratio are considered for this analysis.