Analysis of Long-Period Variable Stars with Nonparametric Tests for Trend Detection

• Activity Number: 573
• Type: Topic Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Woncheol Jang*+ and Cheolwoo Park and Jeongyoun Ahn and Martin Hendry
• Companies: The University of Georgia and The University of Georgia and The University of Georgia and University of Glasgow
• Keywords: Adaptive Neyman test ; Functional clustering; ; Long period variable stars ; Pivotal test ; SiZer analysis ; Trend detection

Abstract

In astronomy the study of variable stars has made crucial contributions to our understanding of many fields, from stellar birth and evolution to the calibration of the extragalactic distance scale. In this paper we perform a time series analysis of the periods between maximum brightness of a group of 378 long period variable stars. The objective of this study is to identify the stars which display certain trends in their periods, via multiple testing of a mean for nonstationary time series model. The novelty of the proposed method is two-fold: first, our approach is not only nonparametric for trends but also for covariance structure of nonstationary noise, for which functional clustering and principal component analysis are used. Another novel feature is the higher power of the proposed test for trend in each time series. The test is based on high dimensional normal mean inference.