Deep learning application for stellar parameters determination: I-constraining the hyperparameters

• Published in: Open astronomy Vol. 31; no. 1; pp. 38 - 57
• Main Authors: Gebran, Marwan, Connick, Kathleen, Farhat, Hikmat, Paletou, Frédéric, Bentley, Ian
• Format: Journal Article
• Language: English
• Published: De Gruyter 17.02.2022
• Subjects: methods: data analysis; methods: deep learning; methods: statistical; Sciences of the Universe; stars: fundamental parameters; techniques: spectroscopic; methods: deep learning; stars: fundamental parameters; methods: statistical; methods: data analysis; techniques: spectroscopic
• ISSN: 2543-6376; 2543-6376
• DOI: 10.1515/astro-2022-0007

Machine learning is an efficient method for analysing and interpreting the increasing amount of astronomical data that are available. In this study, we show a pedagogical approach that should benefit anyone willing to experiment with deep learning techniques in the context of stellar parameter determination. Using the convolutional neural network architecture, we give a step-by-step overview of how to select the optimal parameters for deriving the most accurate values for the stellar parameters of stars: , , [M/H], and . Synthetic spectra with random noise were used to constrain this method and to mimic the observations. We found that each stellar parameter requires a different combination of network hyperparameters and the maximum accuracy reached depends on this combination as well as the signal-to-noise ratio of the observations, and the architecture of the network. We also show that this technique can be applied to other spectral-types in different wavelength ranges after the technique has been optimized.