Detecting multiple periodicities in observational data with the multifrequency periodogram - II. Frequency Decomposer, a parallelized time-series analysis algorithm

摘要

This is a parallelized algorithm performing a decomposition of a noisy timeseries into a number of sinusoidal components. The algorithm analyses allsuspicious periodicities that can be revealed, including the ones that looklike an alias or noise at a glance, but later may prove to be a real variation.After selection of the initial candidates, the algorithm performs a completepass through all their possible combinations and computes the rigorousmultifrequency statistical significance for each such frequency tuple. Thelargest combinations that still survived this thresholding procedure representthe outcome of the analysis. The parallel computing on a graphics processing unit (GPU) is implementedthrough CUDA and brings a significant performance increase. It is stillpossible to run FREDEC solely on CPU in the traditional single-threaded mode,when no suitable GPU device is available. To verify the practical applicability of our algorithm, we apply it to anartificial time series as well as to some real-life exoplanetaryradial-velocity data. We demonstrate that FREDEC can successfully revealseveral known exoplanets. Moreover, it detected a new $9.8$-day variation inthe Lick data for the five-planet system of 55 Cnc. It might indicate theexistence of a small sixth planet in the 3:2 commensurability with the planet55 Cnc b, although this detection is model-dependent and still needs a detailedverification.