Poster abstract details

Periodograms are a family of analytic methods for finding periodic patterns in light curves and other time series. The choice for one periodogram over another generally hinges on its sensitivity (its ability to detect low SNR patterns) and the accuracy of its measured periods. As a general rule, the performance of periodograms can be improved by making assumptions about the periodic patterns one is searching for. However, this comes at a cost of performing poorly in cases where the assumptions do not hold. In this poster I introduce a new class of non-specialized spline-based periodograms that can operate on unevenly sampled observations. These periodograms are a generalization of the Schwarzenberg-Czerny (1989) Analysis of Variance (AoV) periodogram, which in this new framework can be called: 0,0-spline-AoV. When comparing this original AoV periodogram with a 1,1-spline-AoV periodogram, one finds that the new algorithm is slightly more sensitive and about twice as accurate in determining the period of eclipsing binaries in CoRoT-like light curves.

Download poster