Machine Learning methods have emerged as powerful tools for analyzing stellar clusters, which pose significant challenges. techniques such as DBSCAN and GMM have advanced remarkably in this domain. However, these clustering techniques exhibit imperfections and limitations, highlighting the need for careful data tuning and consideration of data characteristics to ensure meaningful result. The utilization of supervised Machine Learning techniques for membership determination of the stellar clusters, especially open clusters, can lead to more accurate results. However, the absence of dataset for training on an open cluster presents a significant hurdle. To address the problem, we’ve introduced a novel approach to generate a labeled dataset for training the super- vised Machine Learning models. Our approach leverages data from Gaia DR3 Catalog, which provides precise astrometric and photometric measurements for millions stars in Milky Way, to construct a comprehensive dataset. Our findings have significant implications for future astronomical research. By using Supervised machine learning techniques, we can achieve more accurate and efficient membership determination for stellar clusters, which can lead to a better understanding of the formation and evolution of galaxies. Our method not only enhances the accuracy of membership determination but also provides insights into the underlying data characteristics that influence cluster analysis.
Read full abstract