Stellar Classification by Machine Learning

Abstract

As an emerging subject with strong comprehensiveness, machine learning has made varying degrees of progress in various fields. In the field of astronomy, it has also been generally used, and there have been quantities of research using machine learning for data processing and model prediction. The paper has used three algorithms (Decision Tree, Random Forest and Support Vector Machine) to build prediction models to classify stars, galaxies, and quasars in the universe and make a comparison among three models. The results of the test have shown that the prediction accuracy of the Random Forest model reaches roughly 98 percent with a great computing efficiency, which performs the best.