Searching for anomalous light curves in massive data sets
Nicholas James Rattenbury · ResearchSpace (University of Auckland) · 2014
The photometric surveys currently under way by the MOA and OGLE collaborations have produced and are extending databases of millions of stellar light curves. These databases have allowed investigations into diverse astrophysical fields including variable stars, proper motion studies and Galactic structure. Odd, or otherwise curious events have been discovered in the databases. We consider here one such event and propose methods for discovering more like it in the microlensing databases. A further aim of this initial work is to set out the prospects of the classification scheme for identifying time series that are maximally discordant - i.e. those that do not look like any other time series in the data set, and which therefore, may be of particular interest.