We describe a method to detect short-term variability based on the change-point analysis with filtering algorithm using local statistics. The use of cumulative sum scheme and bootstrap rank statistics as a means of detecting a series of change points is discussed. By applying this method to over 30,000 lightcurves from the MMT transit survey data, we found previously unknown evidences about stellar variability (including a total of 606 flare events, 18 eclipsing-like features, and 3 transit-like features). In particular, this approach will be effective in detecting non-periodic events in massive astronomical time series data. The detection and characterization of variability is often the first step to understand the nature of various cosmic objects. Most variability detection methods require conventional models that are mainly focused on the strictly periodic signals, and are not suitable for the study of arbitrary-shaped, non-periodic, and sporadically occurring variations, especially those of short time scales. Also, in many cases, signal estimation is equated with smoothing of data for de-noising. This sometimes discards vital information in time series data.We introduce a non-parametric method to extract all significant features based on the change-point analysis (CPA) with filtering algorithm using local statistics.
|Title of host publication||Statistical Challenges in Modern Astronomy V|
|Publisher||Springer Science and Business Media, LLC|
|Number of pages||3|
|Publication status||Published - 2012|
|Event||5th Statistical Challenges in Modern Astronomy Symposium, SCMA 2011 - University Park, PA, United States|
Duration: 2011 Jun 13 → 2011 Jun 15
|Name||Lecture Notes in Statistics|
|Other||5th Statistical Challenges in Modern Astronomy Symposium, SCMA 2011|
|City||University Park, PA|
|Period||11/6/13 → 11/6/15|
Bibliographical noteFunding Information:
This work is supported by Korea Institute of Science and Technology Information under the contract of the commissioned research project, Massive Astronomical Data Applications of Cloud Computation (KISTI-P11020).
All Science Journal Classification (ASJC) codes
- Statistics and Probability
- Statistics, Probability and Uncertainty