Gaia's Source Classifications Compared to NED and SIMBAD
Recent research has compared the classifications of astronomical sources from the Gaia Data Release 3 (DR3) with those from the NASA/IPAC Extragalactic Database (NED) and the SIMBAD astronomical database. The study, titled "Comparing Gaia, NED and SIMBAD source classifications in nearby galaxies," was authored by J. Hales and P. Barmby and was submitted on August 22, 2024.
Gaia DR3 provides classifications for sources in its all-sky database, which includes a significant number of stars from the Milky Way, as well as sources from nearby galaxies and background galaxies and quasars. The researchers matched approximately 320,000 unique Gaia sources with 400,000 sources across 1,040 galaxies in the Local Volume Galaxy catalogue. This matching revealed that the matched sources had a lower fraction of Gaia-classified stars and higher fractions of galaxies and quasars compared to the overall DR3 dataset. Specifically, the matched sources consisted of about 95% stars, 2% galaxies, and 2% quasars.
The study assessed the accuracy of Gaia's classifications by comparing them to the classifications provided by NED and SIMBAD. The balanced accuracy of Gaia's classification was found to be 0.80 when compared to NED and 0.83 when compared to SIMBAD. The most common discrepancies involved literature-classified galaxies being classified as stars by Gaia, and literature-classified stars being classified as quasars.
The purity and completeness metrics indicated that the agreement between Gaia and NED/SIMBAD classifications was highest for stars, with values around 0.9, while the agreement for quasars and galaxies was lower, suggesting challenges in accurately classifying these types of celestial objects. The findings of this research are significant as they highlight the strengths and limitations of the Gaia classification system, which is crucial for astronomers in identifying and studying celestial objects in our universe.
For further details, the paper can be accessed here.