Abstract

We present the physical properties for 2281 northern eclipsing binary (EB) stars with eclipsing Algol (EA)-type light-curve (LC) morphology, based on data extracted from the Catalina Sky Survey (CSS). Our study is based on the analysis of the Eclipsing Binary via Artificial Intelligence (EBAI) artificial neural network (ANN) tool. An intensive search for the optimal ANN topology was performed. In order to feed the ANN with LCs that are representative of the CSS observations, two independent methods, based on template fitting and on the Two-Gaussian Model, were applied. As a result, five principal physical parameters were determined using only the CSS LCs, namely the temperature ratio, T-2/T-1; the sum of relative radii, rho(2) + rho(1); e sin omega; e cos omega; and sin i, where e is the eccentricity, omega is the argument of periastron, and i is the orbital inclination. Parameter uncertainties were estimated based on a Monte Carlo approach. When the ANN predictions were out of its training limits (1540 EBs), the parameters of the systems are based on the matching templates technique only. The results are fully in agreement with the expected parameter values for detached EB systems and can be used as initial inputs for advanced and dedicated EB models and/or for statistical purposes.