CHAMP: Changepoint Detection Using Approximate Model Parameters

Scott Niekum, Sarah Osentoski, Chris Atkeson, and Andrew G. Barto
tech. report CMU-RI-TR-14-10, Robotics Institute, Carnegie Mellon University, June, 2014


Download
  • Adobe portable document format (pdf) (662KB)
Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.

Abstract
We introduce CHAMP, an algorithm for online Bayesian changepoint detection in settings where it is difficult or undesirable to integrate over the parameters of candidate models. Rather than requiring integration of the parameters of candidate models as in several other Bayesian approaches, we require only the ability to fit model parameters to data segments. This approach greatly simplifies the use of Bayesian changepoint detection, allows it to be used with many more types of models, and improves performance when detecting parameter changes within a single model. Experimental analysis compares CHAMP to another state-of-the-art online Bayesian changepoint detection method.

Notes

Text Reference
Scott Niekum, Sarah Osentoski, Chris Atkeson, and Andrew G. Barto, "CHAMP: Changepoint Detection Using Approximate Model Parameters," tech. report CMU-RI-TR-14-10, Robotics Institute, Carnegie Mellon University, June, 2014

BibTeX Reference
@techreport{Niekum_2014_7597,
   author = "Scott Niekum and Sarah Osentoski and Chris Atkeson and Andrew G. Barto",
   title = "CHAMP: Changepoint Detection Using Approximate Model Parameters",
   booktitle = "",
   institution = "Robotics Institute",
   month = "June",
   year = "2014",
   number= "CMU-RI-TR-14-10",
   address= "Pittsburgh, PA",
}