Title: A dynamic Bayesian nonparametric model
Authors: John Maheu - McMaster University (Canada) [presenting]
Yong Song - University of Melbourne (Australia)
Abstract: A Bayesian nonparametric approach is designed to model dynamic changes in an unknown distribution through time. The distribution changes can be interpreted as structure change and we discuss how to perform inference on break dates. A new efficient MCMC routine is provided to estimate the model. Applications to housing data and bank data with comparison to other nonparametric models show the model to work well.