Speaker
Description
The cosmological distribution of galaxies, particularly active ones, serves as a window into the structure and evolution of the universe, and the processes governing galaxy formation, highlighting the co-evolution of galaxies and SMBHs. Although measuring the distances from the spectral observations can consume precious telescope time, predicting photometrical redshift via machine learning algorithms significantly improve precise target selection. Inspired by the studies for high luminosity quasars, we extended the prediction approach for Seyfert galaxies that was relatively less explored.
Employing the optical and infrared colors from SDSS and WISE, we train and test several supervised machine-learning models for approximately 25 000 pre-selected Seyfert galaxies. The best models are Linear regression with R² score of 0.82, Random Forest, and XGBoost models with 0.94. Given the limited redshift range of the Seyfert galaxies, our approach serves as a reliable tool for predicting the distances. Knowing that artificial intelligence has great potential to accelerate astrophysical explorations, our first results prove that a relatively simple application of machine learning can uncover productive uses of precious telescope time.
