Refining the exoplanet mass-radius relation using machine learning

Abstract The rising number of exoplanet discoveries and advances in machine learning (ML) techniques present new possibilities for exploring and understanding the characteristics of worlds beyond our solar system. This research examines the exoplanet dataset by applying ML techniques to categorize these systems, uncover relationships among their physical features, and predict the exoplanet radius. We group the data into two primary categories: ‘small’ and ‘giant’ planets, with thresholds at R p = 8.13 R ⊕ and M p = 52.48 M ⊕ . Our study indicates that the planetary mass, orbital period, and stellar mass play critical roles in predicting the exoplanet radius. A notable finding of our research is that small planets exhibit a positive linear mass-radius relationship, consistent with other studies. Conversely, for giant planets, we observe a strong correlation between planetary radius and the mass of their host stars, potentially providing significant insights into the relationship between giant planet formation and stellar properties.