Quasar and galaxy classification using Gaia EDR3 and CatWise2020

Open Access

Table 1

Classification performances obtained for different balanced classifiers using different algorithms and input features.

		Completeness		Purity

	Features	Star	Quasar	Galaxy	Star	Quasar	Galaxy
GMM	Gaia_f	0.9330	0.9580	0.9886	0.9532	0.9405	0.9860
	Gaia_f + W1-W2	0.9714	0.9850	0.9906	0.9810	0.9784	0.9875
	Gaia_f + W2 + G-W1	0.9766	0.9871	0.9919	0.9853	0.9837	0.9866
	Gaia_f + W1-W2 + G-W1	0.9778	0.9859	0.9919	0.9840	0.9846	0.9869

XGBoost	Gaia_f	0.9418	0.9623	0.9922	0.9603	0.9489	0.9871
	Gaia_f + W1-W2	0.9728	0.9878	0.9933	0.9857	0.9798	0.9885
	Gaia_f + W2 + G-W1	0.9793	0.9896	0.9932	0.9878	0.9859	0.9884
	Gaia_f + W1-W2 + G-W1	0.9793	0.9908	0.9936	0.9891	0.9857	0.9889

CatBoost	Gaia_f	0.9411	0.9619	0.9919	0.9593	0.9484	0.9872
	Gaia_f + W1-W2	0.9720	0.9876	0.9930	0.9854	0.9787	0.9885
	Gaia_f + W2 + G-W1	0.9785	0.9883	0.9927	0.9862	0.9847	0.9886
	Gaia_f + W1-W2 + G-W1	0.9786	0.9905	0.9934	0.9886	0.9850	0.9888

Notes. Completeness and purity are shown for each class. From our tests, the best performing model is the XGBoost algorithm trained on the Gaia_f features supplemented with the infrared CatWise2020 colours (Feature Set 4).

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.