Clustering of eclipsing binary light curves through functional principal component analysis

In this paper, we revisit the problem of clustering 1318 new variable stars found in Milky way. Our recent work distinguishes these based on their light curves which are univariate series brightness from observed at discrete time points. This proposes a approach to look as continuous over by transforming them into functional data. Then, principal component analysis is performed using curves. Clustering significant components reveals two distinct groups eclipsing binaries with consistency and superiority compared our previous results. method established powerful curve-based classifier, where implementation simple algorithm effective enough uncover true clusters merely first few relevant components. Simultaneously discard noise data study involving higher order Thus suggested very useful for big curve sets also verified simulation study.