My latest (Kim+ 2014) work of machine-learning classification for variable star.
We crossmatched EROS-2 counterparts with OGLE-III variables to build a training set, and then used Random Forest to train a classifier.
The classifier applied to the EROS-2 Large Magellanic Cloud containing about 30 million time-series (aka light curves)
The paper presents about 100k new variables.
We also provide a catalog of the variables at:
Example light-curves (phased-folded) of the variable stars classified using our method.
Color-magnitude diagram of the variable stars.
P.S. EPOCH stands for:
EROS-2 PeriOdic variable star Classification using macHine learning
댓글 없음:
댓글 쓰기