Radial velocities data analyzed through artificial neural networks

¹ Laboratoire d'Astrophysique de l'EPFL, Observatoire, 1290 Versoix, Suisse
² Observatoire de Genève, 1290 Versoix, Suisse

Abstract

One of the most efficient methods used for exoplanet detection is the analysis of the radial velocity measurements of the host star, over a period of time. A newly proposed approach to analyze the radial velocity data is based on artificial neural networks (ANNs). The method is inspired from pattern recognition, a domain where ANNs are widely used. Preliminary results show the degree of similarity of the measured radial velocities to the theoretically computed effect of a multi-planetary system on a star. The present analysis can be the starting point for further classification of planetary systems using radial velocity curves.