EPOCH Project. EROS-2 LMC periodic variables

In order to classify these variables, we first build a training set by compiling known variables in the Large Magellanic Could area from the OGLE and MACHO surveys. We crossmatch these variables with the EROS-2 sources and extract 22 variability features from 28,392 light curves of the corresponding EROS-2 sources.

We then use Random Forests to classify the EROS-2 sources in the training set. We design the model to separate not only Delta Scuti stars, RR Lyraes, Cepheids, eclipsing binaries and long-period variables, the "superclasses", but also their subclasses, such as RRab, RRc, RRd and RRe for RR Lyraes, and similarly for the other variable types. We apply the trained model to the entire EROS-2 LMC database containing about 29 million sources and find 117,234 periodic variable candidates.