Fundamental Parameter Estimations of O-type Stars Binary Systems Using Recurrent Neural Networks
Recent studies have established O-type stars as predominantly born in binary star systems. In this work, we explore the application of a recurrent neural network system for estimating the effective temperature and surface gravity of binary star systems. Additionally, we assess the neural network's sensitivity in processing synthetic binary spectra derived from two stellar spectra models of O-type stars and the implications of the contribution from the secondary star in the system. Finally, we compare the estimations produced by our proposed system with those from prior research.
Miguel Flores R. is a Ph.D. student at the University of Guadalajara working on the applications of ANN for processing and fitting models of stellar spectra. He is also a Sr. Data Science Globant developing models for systems that go from optimization in the energy and natural resources sector, to gaming live operations. Flores holds a BSc and MSc degrees in physics from the University of Guadalajara. During his former years, his research interests were in the areas of quantum information theory and complex systems focused on discrete systems dynamics. In 2018 has the opportunity of working in the Global IA and Data Science team at HP Inc. in Palo Alto, CA, working on human behavior models and web interactions. His current research project explores the multidisciplinary application of IA structures to process, analyze and model big dimensional systems.