VizieR Online Data Catalog: KiDSDR4 QSOs photometric redshifts catalog (Nakoneczny+, 2021)

The catalog results from applying artificial neural networks to process the KiDS data limited to 9-band detections. The machine learning (ML) models are trained on KiDS objects cross-matched with the spectroscopic SDSS survey. We address the problem of extrapolation to KiDS objects fainter than the SDSS limit by properly generalising ML models, and creating inference subsets which describe the reliability of estimations: safe, extrapolation, unsafe. We provide the suggested cuts on magnitude and probability of photometric classification, which are derived from validating the catalog with several methods.

(2 data files).