A Deep Learning Neural Network Algorithm for Classification of Eclipsing Binary Light Curves

Autor:	Ulas, Burak
Rok vydání:	2023
Předmět:	Astrophysics - Solar and Stellar Astrophysics Astrophysics - Instrumentation and Methods for Astrophysics
Druh dokumentu:	Working Paper
Popis:	We present an image classification algorithm using deep learning convolutional neural network architecture, which classifies the morphologies of eclipsing binary systems based on their light curves. The algorithm trains the machine with light curve images generated from the observational data of eclipsing binary stars in contact, detached and semi-detached morphologies, whose light curves are provided by Kepler, ASAS and CALEB catalogues. The structure of the architecture is explained, the parameters of the network layers and the resulting metrics are discussed. Our results show that the algorithm, which is selected among 132 neural network architectures, estimates the morphological classes of an independent validation dataset, 705 real data, with an accuracy of 92%. Comment: 22 pages, 10 figures
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2306.02686 Zobrazit plný text záznamu View this record from Arxiv