Automated Classification of Celestial Objects Using Machine Learning

The swift expansion of astronomical data requires the automated classification of celestial objects for practical use. Because of its manual and monotonous nature, classification is more prone to errors and is rapidly losing its viability. This study performs the classification of stars, galaxies, and quasars from SDSS (Sloan Digital Sky Survey) data using the Random Forest, XGBoost, Decision Tree, Gradient Boosting, Linear SVM, KNN, and Logistic Regression. In order to fix the imbalance in the data, the SMOTE algorithm was used, making the model more robust. Random Forest topped the models with their accuracy and reliability across many multiple data releases, hitting an astonishing 99.12% accuracy in SDSS DR18. This work shows how much machine deep learning can change astronomical surveys, providing readily available, precise techniques that are much more effective than manual approaches. The results add towards the development of astrophysics while simultaneously meeting Sustainable Development Goal 9 on fostering innovation through the need for infrastructure