A0925
Title: Joint random partition models for multivariate change point analysis
Authors: Garritt Page - Brigham Young University (United States) [presenting]
Jose Javier Quinlan Binelli - Pontificia Universidad Catolica de Chile (Chile)
Mauricio Castro - Pontificia Universidad Catolica de Chile (Chile)
Abstract: Change point analyses identify positions of an ordered stochastic process that undergo abrupt local changes of some underlying distribution. When multiple procedures are observed, information regarding the change point positions is often shared across the different approaches. A method that takes advantage of this type of information is described. Since the number and position of change points can be described through a partition with contiguous clusters, this approach develops a joint model for these types of partitions. Computational strategies associated with our approach are described, and improved performance in detecting change points through a small simulation study is illustrated. This method is applied to a financial data set of emerging markets in Latin America, and interesting insights discovered due to the correlation between change point locations among these economies are highlighted.
