Estimation of zero-inflated parameter-driven models via data cloning

ABSTRACT

In this paper, we propose the use of the Data Cloning (DC) approach to estimate parameter-driven zero-inflated Poisson and Negative Binomial models for time series of counts. The data cloning algorithm obtains the familiar maximum likelihood estimators and their standard errors via a fully Bayesian estimation. This provides some computational ease as well as inferential tools such as confidence intervals and diagnostic methods which, otherwise, are not readily available for parameter-driven models. To illustrate the performance of the proposed method, we use Monte Carlo Simulations and real data on asthma-related emergency department visits in the Canadian province of Ontario.

KEYWORDS:

• Zero-inflation
• parameter-driven
• Poisson
• negative binomial
• Bayesian estimation
• state-space models
• data cloning
• time series of counts

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The authors would like to thank Natural Sciences and Engineering Research Council of Canada (NSERC) discovery grant for supporting this research.