Venkatasubramaniam, A., Evers, L. and Ampountolas, K. (2016) Distance Dependent Chinese restaurant process for Spatio-Temporal Clustering of Urban Traffic Networks. 22nd International Conference on Computational Statistics (COMPSTAT 2016), Oviedo, Spain, 23-26 Aug 2016.
Full text not currently available from Enlighten.
Publisher's URL: http://cmstatistics.org/RegistrationsV2/COMPSTAT2016/viewSubmission.php?in=511&token=n6n7ss8613qp3n67o627201p5soqr280
Abstract
A novel Bayesian clustering method is presented for spatio-temporal data observed on a network. This method employs a Distance Dependent Chinese Restaurant Process (DDCRP) to incorporate the geographic constraints of the network. DDCRPs typically accommodate non-exchangeable distributions as a prior over partitions unlike the traditional Chinese restaurant Process. In addition, it does not exhibit the marginal invariance property and so one can capture the extent of the influence transferred from one node in the network to the next. We do not expect the DDCRP to fully capture the dependency structure of the data and thus a conditional auto-regressive model (CAR) is used to model the spatial dependency within a cluster. We will discuss different strategies for incorporating temporal dependency into a CAR-type model. Inference is carried out using a Metropolis within Gibbs sampler and we apply the model to cluster an urban traffic network, using occupancy data recorded at the junction level.
Item Type: | Conference or Workshop Item |
---|---|
Keywords: | Distance dependent Chinese restaurant process, patio-temporal clustering, urban road networks |
Status: | Published |
Refereed: | Yes |
Glasgow Author(s) Enlighten ID: | Ampountolas, Dr Konstantinos and Evers, Dr Ludger and Venkatasubramaniam, Ashwini |
Authors: | Venkatasubramaniam, A., Evers, L., and Ampountolas, K. |
College/School: | College of Science and Engineering College of Science and Engineering > School of Engineering > Infrastructure and Environment College of Science and Engineering > School of Mathematics and Statistics > Statistics |
Related URLs: |
University Staff: Request a correction | Enlighten Editors: Update this record