THE COROT PUBLIC ARCHIVE AT CAB
COROT VARIABILITY CLASSES
Notes on the CoRoT variability classifier:
- The class codes (columns 'Var1-3') and corresponding probabilities (columns 'Prob1-3')
can be used to make candidate lists with different levels of contamination. For example, users interested in RR-Lyrae
stars of subtype ab, and wishing to minimize the risk of missing any, will look for those light curves having one of
the 3 class codes equal to 'RRAB'.
- Minimizing the risk of missing any objects of that type unfortunately also implies that the candidate list
will be more contaminated by false-positives (note that also non-variable light curves get class probabilities).
Stronger selections can be made by selecting only those objects having 'Var1' or 'Var2' of the desired type, or only
those having 'Var1' of the desired type.
- Imposing limits on the class probabilities will further clean the selection. In general, the higher the
respective class probability the better the candidate and the more similar it is to the objects used to define the
class. However, low class probabilities (even lower than 0.5) do not necessarily imply that we are dealing with a bad
candidate. BCEP and DSCUT stars for example can show very similar pulsation characteristics. Remember that the
current classifiers do not take colour information into account, which would allow for a better separation of these 2
classes. Hence, the light curve of a BCEP star might get similar probabilities for the BCEP and DSCUT classes.
The same is true for the SPB and GDOR classes, they are also difficult to distinguish without any spectral
- Note that the listed probabilities are relative ones: all the class probabilities for every light curve sum
up to 1. In the classification file, we only list the 3 most probable classes, in order of decreasing probability.
- More detailed information on the classification methods can be found in the following publications:
- Debosscher et al., 2009 (A&A): Automated supervised classification of variable stars in the CoRoT programme.
- Debosscher et al., 2007 (A&A): Automated supervised classification of variable stars. I. Methodology.
- Sarro et al., 2009 (A&A): Automated supervised classification of variable stars. II. Application to the OGLE database.
- Sarro et al., 2009 (A&A): Comparative clustering analysis of variable stars in the Hipparcos, OGLE Large Magellanic Cloud, and CoRoT exoplanet databases.