Press Release
Open Access

Table 6.

Resulting metrics of the radio detection model on the test subset and the labelled sources in S82 using two different threshold values, as explained in Sect. 4.2.

Subset Threshold Fβ MCC Precision Recall
(× 100) (× 100) (× 100) (× 100)
HETDEX-test Naive 24.87 ± 2.94 27.36 ± 3.46 60.61 ± 8.18 16.72 ± 2.31
PR 42.88 ± 2.93 32.47 ± 3.49 35.28 ± 2.74 52.16 ± 3.59
S82-label Naive 27.15 ± 2.28 23.36 ± 2.27 25.72 ± 1.91 28.47 ± 3.24
PR 21.62 ± 1.20 19.37 ± 1.64 12.29 ± 0.73 58.16 ± 3.06
HETDEX-pipe Naive 24.37 ± 3.53 26.93 ± 4.18 59.36 ± 7.17 16.38 ± 2.63
PR 41.57 ± 4.16 31.67 ± 4.81 34.65 ± 3.24 49.80 ± 5.85
S82-pipe Naive 26.52 ± 5.44 23.29 ± 5.73 25.71 ± 5.89 27.72 ± 5.21
PR 20.19 ± 2.84 18.40 ± 4.07 11.45 ± 1.58 54.78 ± 8.44

Notes. HETDEX and S82 pipeline results shown as part of the discussion in Sect. 4.4. Values and uncertainties as in Table 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.