Using Machine Learning to Estimate the Photometric Redshift of Galaxies
Date
2023-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of the Witwatersrand, Johannesburg
Abstract
Machine learning has emerged as a crucial tool in the field of cosmology and astrophysics, leading to extensive research in this area. This research study aims to utilize machine learning models to estimate the redshift of galaxies, with a primary focus on utilizing photometric data to obtain accurate results. Five machine learning algorithms, including XGBoost, Random Forests, K-nearest neighbors, Artificial Neural Networks, and Polynomial Regression, are employed to estimate the redshifts, trained on photometric data derived from the Sloan Digital Sky Survey (SDSS) Data Release 17 database. Furthermore, various input parameters from the SDSS database are explored to achieve the most accurate redshift values. The research incorporates a comparative analysis, utilizing different evaluation metrics and statistical tests to determine the best-performing algorithm. The results indicate that the XGBoost algorithm achieves the highest accuracy, with an R2 value of 0.94, a Root Mean Square Error (RMSE) of 0.03, and a Mean Absolute Average Percentage (MAPE) of 12.04% when trained on the optimal feature subset. In comparison, the base model achieved an R2 of 0.84, a RMSE of 0.05, and a MAPE of 20.89%. The study contributes to the existing literature by utilizing photometric data during model training and comparing different high-performing algorithms from the literature.
Description
A research report submitted in partial fulfillment of the requirements for the degree of Master of Science (Artificial Intelligence), to the Faculty of Science, the School of Computer Science and Applied Mathematics, University of the Witwatersrand, Johannesburg, 2023.
Keywords
Machine Learning, Genetic Algorithms, Photometric Redshift Estimation, Astrophysics, UCTD
Citation
Salim, Shayaan. (2023). Using Machine Learning to Estimate the Photometric Redshift of Galaxies. [Master's dissertation, University of the Witwatersrand, Johannesburg]. https://hdl.handle.net/10539/41702