DeepAI AI Chat
Log In Sign Up

Warped Dynamic Linear Models for Time Series of Counts

by   Brian King, et al.
Rice University

Dynamic Linear Models (DLMs) are commonly employed for time series analysis due to their versatile structure, simple recursive updating, and probabilistic forecasting. However, the options for count time series are limited: Gaussian DLMs require continuous data, while Poisson-based alternatives often lack sufficient modeling flexibility. We introduce a novel methodology for count time series by warping a Gaussian DLM. The warping function has two components: a transformation operator that provides distributional flexibility and a rounding operator that ensures the correct support for the discrete data-generating process. Importantly, we develop conjugate inference for the warped DLM, which enables analytic and recursive updates for the state space filtering and smoothing distributions. We leverage these results to produce customized and efficient computing strategies for inference and forecasting, including Monte Carlo simulation for offline analysis and an optimal particle filter for online inference. This framework unifies and extends a variety of discrete time series models and is valid for natural counts, rounded values, and multivariate observations. Simulation studies illustrate the excellent forecasting capabilities of the warped DLM. The proposed approach is applied to a multivariate time series of daily overdose counts and demonstrates both modeling and computational successes.


page 1

page 2

page 3

page 4


Latent Gaussian dynamic factor modeling and forecasting for multivariate count time series

This work considers estimation and forecasting in a multivariate count t...

Bayesian forecasting of multivariate time series: Scalability, structure uncertainty and decisions

I overview recent research advances in Bayesian state-space modeling of ...

Bayesian time-aligned factor analysis of paired multivariate time series

Many modern data sets require inference methods that can estimate the sh...

Bayesian Computation in Dynamic Latent Factor Models

Bayesian computation for filtering and forecasting analysis is developed...

Learning a latent pattern of heterogeneity in the innovation rates of a time series of counts

We develop a Bayesian hierarchical semiparametric model for phenomena re...

Panel data analysis via mechanistic models

Panel data, also known as longitudinal data, consist of a collection of ...

A computational scheme for Reasoning in Dynamic Probabilistic Networks

A computational scheme for reasoning about dynamic systems using (causal...