A new approach based on AI tools for mass-radius ratio in exoplanets

Joaquín Álvaro,¹ Carlos D. Álvaro ²

 ¹ Europlanet Society - Spain and Portugal Regional Hub
² Meteologica - Global Weather Service Provider

ABSTRACT

The problem derived from the lack of observational measures for the masses and radios in the set of confirmed exoplanets and also of possible candidates is known. In almost 80%, one or the other of these parameters is known depending on the method of discovery, RV or Transit, but not both at the same time. This is a manifest difficulty for its characterization. Until now, the probabilistic forecast provided by the algorithm of Chen and Kipping (2017), based on a Bayesian statistic, is assumed as a reasonable compromise solution and also covers a wide range of planetary sizes. However, we consider that the scenario in the confluence region of rocky planets, Earth type, and gas, Neptune type, is too simple and ignores details that give this area a significantly richer and more complex structure. We propose here an alternative model based on the application of machine-learning tools on the Composite Planet Data (NASA Exoplanet Archive) for the characterization of exoplanets with R < 5 R_⊕.

LEER MÁS ...