Identifying radio active galactic nuclei with machine learning and large-area surveys

Home

All issues

Volume 701 (September 2025)

A&A, 701 (2025) A179

Full HTML

Open Access

Table 4.

Performance for different combinations of training features.

Features	AGN			non-AGN
Features	Precision	Recall	F₁ score	Precision	Recall	F₁ score
W1, W2, W3, W4	0.978	0.794	0.877	0.906	0.991	0.946
W1 − W2, W2 − W3, W3 − W4	0.943	0.776	0.852	0.897	0.976	0.935
W1, W2, W3, W4, W1 − W2, W2 − W3, W3 − W4	0.965	0.812	0.882	0.913	0.985	0.948
G, BP, RP, BP − G, G − RP	0.901	0.747	0.817	0.883	0.959	0.919
W1 − W2, W2 − W3, W3 − W4, BP − G, G − RP	0.958	0.812	0.879	0.912	0.982	0.946
W1, W2, W3, W4, G, BP, RP	0.966	0.824	0.889	0.918	0.985	0.950
W1, W2, W3, W4, W1 − W2, W2 − W3, W3 − W4,	0.973	0.853	0.910	0.931	0.988	0.959
G, BP, RP, BP − G, G − RP, F_total	0.973	0.853	0.910	0.931	0.988	0.959
W1, W2, W3, W4, W1 − W2, W2 − W3, W3 − W4,	0.973	0.841	0.902	0.925	0.988	0.956
G, BP, RP, BP − G, G − RP, F_total/F_W3	0.973	0.841	0.902	0.925	0.988	0.956
W1, W2, W3, W4, W1 − W2, W2 − W3, W3 − W4,	0.963	0.949	0.956	0.816	0.861	0.838
G, BP, RP, BP − G, G − RP, parallax, propermotion*	0.963	0.949	0.956	0.816	0.861	0.838

Notes. The number of AGNs and non-AGNs containing measurements of parallax and proper motion are 455 and 116, respectively.

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.