Light Curve Analysis for Classification with Astronomical Data

• Activity Number: 415
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Ashish Mahabal*+ and Julian Faraway and Jiayang Sun and Xiaofeng Wang and Yi Wang and Lingsong Zhang
• Companies: Caltech and Department of Mathematical Sciences, University of Bath and Case Western Reserve University and Cleveland Clinic Lerner Research Institute and SAMSI/Duke University and Purdue University
• Keywords: astronomy ; images ; time series

Abstract

Clustering and Classification are extremely important in astronomy: right from separating stars and galaxies in large sky surveys, to identifying rare transients in real-time from synoptic surveys. In this talk, we present light curves and other image-based parameters crucial for this process. A majority of the objects we deal with are transients, but also included are some strong variables which one has to study in conjunction with transients for successful separation of differnet types of transients. Some of the parameters come with associated measurement errors, and partial knowledge of the membership of the objects is available a priori. We develop a new semi-supervised procedure that is effective in classifying the objects and compares well with existing methods. Finally we detail the significance of this to the large astronomical surveys of the near future.