Multi-Algorithm Analysis Of The Semi-Regular Variable DY Per, The Prototype Of The Class Of Cool RCrB Variables

TL;DR

This study compares multiple time series analysis algorithms applied to the semi-regular variable DY Per, revealing a change in its photometric period and introducing a sinusoidality parameter to characterize its variability.

Contribution

It provides a comparative analysis of 21 phenomenological algorithms for variable star analysis and applies them to DY Per, highlighting period changes and introducing a new variability parameter.

Findings

Photometric period changed from 851.1d to 780.5d.

Introduced a sinusoidality parameter for variability characterization.

Demonstrated effectiveness of phenomenological algorithms in variable star analysis.

Abstract

Multiple algorithms of time series analysis are briefly reviewed and partially illustrated by application to the visual observations of the semi-regular variable DY Per from the AFOEV database. These algorithms were implemented in the software MCV (Andronov and Baklanov, 2004), MAVKA (Andrych and Andronov, 2019; Andrych et al., 2019). Contrary to the methods of physical modeling, which need to use too many parameters, many of which may not be determined from pure photometry (like temperature/spectral class, radial velocities, mass ratio), phenomenological algorithms use smaller number of parameters. Beyond the classical algebraic polynomials, in the software MAVKA are implemented other algorithms, totally 21 approximations from 11 classes. Photometric observations of DY Per from the AFOEV international database were analyzed. The photometric period has switched from P=851.1d(4.1) to P=780.5d(2.7) after JD 2454187(9)d. A parameter of sinusoidality is introduced, which is equal to the ratio of effective semi-amplitudes of the signal determined from a sine fit and the running parabola scalegram.