Improving the open cluster census - I. Comparison of clustering algorithms applied to Gaia DR2 data

Free Access

Table 4.

Extra information on the algorithms’ performance.

Algorithm	Reported OC candidates ^(a)	Fraction with CST > 3σ	Total crossmatches ^(b)	Mean runtime (mins) ^(c)
DBSCAN (ACG)	1518–1538	58.9%–59.6%	382	1.19 (1 repeat)–10.3 (30 repeats)
DBSCAN (model)	5212–51920	22.4%–2.1%	593	0.885
HDSBCAN	1196–49693	82.0%–5.2%	756	2.36
GMM	314–2465	60.5%–20.5%	213	21.9 (m_s = 800) 47.0 (variable m_s)

Notes.

^(a)

Total number of OC candidates for all fields that passed the proper motion dispersion and radius constraints from Sect. 4.2. The range is between the minimum and maximum reported number for the least and most sensitive parameters.

^(b)

Total crossmatches is given as the union of all results from all parameter sets for a given method. A total of 1385 literature OCs were crossmatched against.

^(c)

Mean runtime for a single field out of the 100 in this study. All runs were conducted on the same workbench computer with a 3.1 GHz 4-core CPU.

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.