Zero-Inflated Power Series Joint Model to Analyze Count Data with Missing Responses

Abstract

A zero-inflated power series (ZIPS) joint model for count data with excess zeros and missing responses is presented. A full likelihood-based approach is implemented to obtain maximum likelihood estimates of the model parameters. The purpose of this article is to propose the mixed SEM and EM algorithms (M-SEM-EM algorithm) for parameter estimation in the likelihood function. To illustrate the utility of the proposed model, a large data set excerpted from the British Household Panel Survey (BHPS) is analyzed.

Keywords:

• 62P99

Keywords:

• EM algorithm
• Excess zeros
• Generalized linear mixed model
• Missing Responses
• Stochastic EM algorithm
• Zero-inflated model