Machine Learning Algorithms for Predicting Stellar Light Curves of RR Lyrae Variables

Abstract

The question of why stars are dim and bright has been an important one within the physics realm for decades. This question is thoroughly answered within this paper, which aims to investigate and successfully predict the stellar light curve values, measured in magnitude, of the star, RR Lyrae. The prediction of RR Lyrae light curves are crucial and relevant to astrophysical research, as an accurate prediction allows insight to factors such as cosmic distances. To do so, various computation methods are employed, utilizing Python’s Matplotlib for analysis and visualization, specifically consisting of basic mathematical regression models, machine learning algorithms, and neural networks. Hence, my study also aims to highlight the capabilities and limitations of each method, which provides insight into their performance. This will be furthermore measured with evaluation metrics, consisting of mean squared error and mean absolute error, respectively. Thus, the findings from this study will provide insight into the prediction of RR Lyrae’s stellar light curve values, and emphasize the importance of selecting appropriate techniques for precise predictions. Moreover, the findings also illustrate the capabilities of various machine learning algorithms within different scenarios and contexts, shedding light into the inherent features and advantages of each one.