Week 7: Machine Learning Basics for Astronomy

In this session, we look at the fundamental concepts of machine learning with a specific focus on their applications in astronomy. We cover essential topics, including supervised learning for classification, where you'll learn to classify astronomical objects using real datasets from the SDSS (Sloan Digital Sky Survey) DR16. We'll walk you through data acquisition from VizieR, exploratory data analysis, handling data imbalance, and training various classifiers like Random Forest and Logistic Regression. We also explore Principal Component Analysis (PCA) for dimensionality reduction, demonstrating how this powerful technique can help reduce the complexity of astronomical datasets by identifying key components, with applications to simulated astronomical spectra and a discussion on the importance of data scaling. Finally, we introduce you to the AstroML package, a robust Python library built specifically for machine learning in astronomy. We'll show you how to access its resources and leverage Google Colab's AI features to enhance your projects. This session provides hands-on experience and valuable insights for anyone looking to apply machine learning techniques to astronomical data. We also offer guidance for your Capstone project, encouraging you to explore AstroML notebooks for inspiration and to clearly define your research questions. Don't forget to like, share, and subscribe for more content from BRICS Astronomy!