Free Access

Table 5

Recall and precision of the random forest model considering subclasses.

Class Subclass Recall Precision
(%) (%)

DSCT 97.6 98.0
CEPH
F 97.3 97.7
1O 91.1 91.6
Other 71.3 73.1
RRL
ab 99.6 99.0
c 95.9 87.4
d 65.1 93.3
e 83.1 85.5
EB
EC 52.6 70.1
ED 88.9 79.9
ESD 73.0 65.0
ED+ESD 5.6 50.0
Other 11.0 44.4
LPV
Mira AGB C 88.3 86.5
Mira AGB O 82.1 86.9
OSARG AGB C 57.0 71.9
OSARG AGB O 80.8 76.1
OSARG RGB C 13.0 30.0
OSARG RGB O 84.4 83.4
SRV AGB C 88.2 85.3
SRV AGB O 84.4 84.9
T2CEPH 71.4 88.9
BV 96.1 97.2
QSO 92.5 91.4
NonVar 98.2 97.6

Average recall and precisiona 86.7 86.3

Notes.

(a)

Average weighted by the proportion of classes in the training samples.

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.