Application of the Trend Filtering Algorithm in the search for multiperiodic signals

TL;DR

This paper extends the Trend Filtering Algorithm (TFA) to detect multiperiodic signals in photometric data, demonstrating improved detection efficiency and signal reconstruction, with applications to HATNet survey data.

Contribution

The paper introduces an extension of TFA for multiperiodic signals and shows its effectiveness over standard methods in photometric variability analysis.

Findings

High efficiency in detecting multiperiodic signals.

Incidence rates of 4-10% for variable stars.

Many variables with sub-millimagnitude amplitudes.

Abstract

During the past few years the Trend Filtering Algorithm (TFA) has become an important utility in filtering out time-dependent systematic effects in photometric databases for extrasolar planetary transit search. Here we present the extension of the method to multiperiodic signals and show the high efficiency of the signal detection over the direct frequency analysis on the original database derived by today's standard methods (e.g., aperture photometry). We also consider the (iterative) signal reconstruction that involves the proper extraction of the systematics. The method is demonstrated on the database of fields observed by the HATNet project. A preliminary variability statistics suggests incidence rates between 4 and 10% with many (sub)mmag amplitude variables.