-
Automatic classification of eclipsing binary stars using deep learning methods
Authors:
Michal Čokina,
Viera Maslej-Krešňáková,
Peter Butka,
Štefan Parimucha
Abstract:
In the last couple of decades, tremendous progress has been achieved in develo** robotic telescopes and, as a result, sky surveys (both terrestrial and space) have become the source of a substantial amount of new observational data. These data contain a lot of information about binary stars, hidden in their light curves. With the huge amount of astronomical data gathered, it is not reasonable to…
▽ More
In the last couple of decades, tremendous progress has been achieved in develo** robotic telescopes and, as a result, sky surveys (both terrestrial and space) have become the source of a substantial amount of new observational data. These data contain a lot of information about binary stars, hidden in their light curves. With the huge amount of astronomical data gathered, it is not reasonable to expect all the data to be manually processed and analyzed. Therefore, in this paper, we focus on the automatic classification of eclipsing binary stars using deep learning methods. Our classifier provides a tool for the categorization of light curves of binary stars into two classes: detached and over-contact. We used the ELISa software to obtain synthetic data, which we then used for the training of the classifier. For evaluation purposes, we collected 100 light curves of observed binary stars, in order to evaluate a number of classifiers. We evaluated semi-detached eclipsing binary stars as detached. The best-performing classifier combines bidirectional Long Short-Term Memory (LSTM) and a one-dimensional convolutional neural network, which achieved 98% accuracy on the evaluation set. Omitting semi-detached eclipsing binary stars, we could obtain 100% accuracy in classification.
△ Less
Submitted 3 August, 2021;
originally announced August 2021.
-
ELISa: A new tool for fast modelling of eclipsing binaries
Authors:
Michal Čokina,
Miroslav Fedurco,
Štefan Parimucha
Abstract:
We present a new, fast, and easy to use tool for modelling light and radial velocity curves of close eclipsing binaries with built-in methods for solving an inverse problem. The main goal of ELISa (Eclipsing binary Learning and Interactive System) is to provide an acceptable compromise between computational speed and precision during the fitting of light curves and radial velocities of eclipsing b…
▽ More
We present a new, fast, and easy to use tool for modelling light and radial velocity curves of close eclipsing binaries with built-in methods for solving an inverse problem. The main goal of ELISa (Eclipsing binary Learning and Interactive System) is to provide an acceptable compromise between computational speed and precision during the fitting of light curves and radial velocities of eclipsing binaries. The package is entirely written in the Python programming language in a modular fashion, making it easy to install, modify, and run on various operating systems. ELISa implements Roche geometry and the triangulation process to model a surface of the eclipsing binary components, where the surface parameters of each surface element are treated separately. Surface symmetries and approximations based on the similarity between surface geometries were used to reduce the runtime during light curve calculation significantly. ELISa implements the least square trust region reflective algorithm and Markov-chain Monte Carlo optimisation methods to provide the built-in capability to determine parameters of the binary system from photometric observations and radial velocities. The precision and speed of the light curve generator were evaluated using various benchmarks. We conclude that ELISa maintains an acceptable level of accuracy to analyse data from ground-based and space-based observations, and it provides a significant reduction in computational time compared to the current widely used tools for modelling eclipsing binaries.
△ Less
Submitted 18 June, 2021;
originally announced June 2021.
-
Optical flare activity in the low-mass eclipsing binary GJ~3236
Authors:
S. Parimucha,
P. Dubovsky,
M. Vanko,
M. Cokina
Abstract:
We present our observations of the low-mass eclipsing binary GJ~3236. We have analyzed a phased $R_C$ light-curve and confirmed previously determined fundamental parameters of the components. We detected evolution of the spot(s) and found that there exists a large spot near a polar region of the primary component and another spot either on the primary or the secondary component. We also observed 7…
▽ More
We present our observations of the low-mass eclipsing binary GJ~3236. We have analyzed a phased $R_C$ light-curve and confirmed previously determined fundamental parameters of the components. We detected evolution of the spot(s) and found that there exists a large spot near a polar region of the primary component and another spot either on the primary or the secondary component. We also observed 7 flare events and determined a flare rate of about 0.1 flares per hour. We observed two high energy, long-term flares with a complex light curve and possibly four weak short-term flaring events. A majority of the flares was detected in the $R_C$ filter, which indicate their high energy.
△ Less
Submitted 3 August, 2016;
originally announced August 2016.
