Generalized Polya Urn for Time-varying Dirichlet Process Mixtures
arXiv:1206.5254
Abstract
Dirichlet Process Mixtures (DPMs) are a popular class of statistical models to perform density estimation and clustering. However, when the data available have a distribution evolving over time, such models are inadequate. We introduce here a class of time-varying DPMs which ensures that at each time step the random distribution follows a DPM model. Our model relies on an intuitive and simple generalized Polya urn scheme. Inference is performed using Markov chain Monte Carlo and Sequential Monte Carlo. We demonstrate our model on various applications.
Appears in Proceedings of the Twenty-Third Conference on Uncertainty in Artificial Intelligence (UAI2007)
Cited by in corpus (19)
- Are Gibbs-type priors the most natural generalization of the Dirichlet process?
- Bayesian nonparametric forecasting of monotonic functional time series
- Dynamic density estimation with diffusive Dirichlet mixtures
- Conjugacy properties of time-evolving Dirichlet and gamma random measures
- Bayesian nonparametric dependent model for partially replicated data: the influence of fuel spills on species diversity
- A unifying representation for a class of dependent random measures
- Inference for multiple object tracking: A Bayesian nonparametric approach
- Time-dependent Hierarchical Dirichlet Model for Timeline Generation
- Distance Dependent Infinite Latent Feature Models
- Scalable Bayesian reduced-order models for high-dimensional multiscale dynamical systems
- Using informative priors in the estimation of mixtures over time with application to aerosol particle size distributions
- A Nonparametric Bayesian Model for Sparse Dynamic Multigraphs
- Applications of Probabilistic Programming (Master's thesis, 2015)
- A New Class of Time Dependent Latent Factor Models with Applications
- General Bayesian inference schemes in infinite mixture models
- Unsupervised Detection and Tracking of Arbitrary Objects with Dependent Dirichlet Process Mixtures
- Flexible online multivariate regression with variational Bayes and the matrix-variate Dirichlet process
- Multilinear Dirichlet Processes
- A survey of non-exchangeable priors for Bayesian nonparametric models