QSOs Selection in Highly Unbalanced Photometric Datasets: The “Michelangelo” Reverse-Selection Method

We present a novel selection method aimed at efficiently identifying high-redshift QSOs in highly unbalanced photometric datasets. The method relies on a gradient boosting algorithm, although it may be be used with any other machine learning method providing classification probabilities. We applied the selection method on a photometric dataset and compared its performances to its direct-selection method counterpart, showing that the former privileges the completeness, while the latter privileges the success rate.