Sparse Box-fitting Least Squares

TL;DR

The paper introduces Sparse BLS, an improved algorithm for detecting transiting exoplanets that is more efficient and phase-independent than traditional BLS, especially suitable for large, sparse photometric datasets.

Contribution

It presents Sparse BLS, a novel implementation that eliminates data binning and phase grid reliance, enhancing detection efficiency and speed for unevenly-sampled data.

Findings

Detection efficiency is slightly better than BLS.

Significantly faster for sparse data.

Suitable for large photometric surveys like Gaia.

Abstract

We present a new implementation of the commonly used Box-fitting Least Squares (BLS) algorithm, for the detection of transiting exoplanets in photometric data. Unlike BLS, our new implementation - Sparse BLS (SBLS), does not use binning of the data into phase bins, nor does it use any kind of phase grid. Thus, its detection efficiency does not depend on the transit phase, and is therefore slightly better than that of BLS. For sparse data, it is also significantly faster than BLS. It is therefore perfectly suitable for large photometric surveys producing unevenly-sampled sparse light curves, such as Gaia.