Bayesian Zero-Inflated Poisson Model and Its Application in Age-Specific Fertility Rate Pattern Estimation

• Activity Number: 520
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Yi Cao*+ and Jerome Keating and Keying Ye
• Companies: The University of Texas at San Antonio and The University of Texas at San Antonio and The University of Texas at San Antonio
• Keywords: Zero-inflated Poisson ; Age-specific Fertility Rate ; Metropolis-Hastings ; Excess zeros

Abstract

In this research, we take into account the presence of excess zeros in fertility pattern estimation and propose a zero-inflated Poisson (ZIP) model with skewed-normal (SN) function added as its rate trajectory. The hierarchical structure stabilizes the estimation of zero-inflated probabilities at different childbearing ages.

In terms of computation, the Metropolis-Hasting algorithm has been applied to estimate parameters in the model and curve precedence test is adopted to test the equality of multiple curves which makes the conclusion statistical intact.

The practicability of the ZIP-SN model is illustrated with a real-life data set of fertility numbers of 254 counties in Texas over six years from 2000 to 2005. Our model delivers not only sound fitness for the data set but also offers more rational statistical explanation than traditional age-specific fertility rate pattern estimation methods and empirical calculation method.