ON ESTIMATION AND TESTS FOR ZERO INFLATED REGRESSION MODELS

by Sumathi K. and Aruna Rao K.

Abstract: The zero altered count models are being widely used in various disciplines such as econometrics and medicine. Many of the applications involve the zero inflated Poisson (ZIP) model and its generalizations. A few applications can also be found using zero altered negative binomial and binomial distributions. This paper provides a brief review of the literature on zero inflated models (ZIM). The regularity conditions for the existence of maximum likelihood (ML) estimators and the computational difficulties in solving the ML equations are discussed. The theoretical issues concerning this are discussed using a data set from dental epidemiological study.

Key Words: zero inflated distributions, regression model, maximum likelihood estimators, optimal equations, dental epidemiology

Authors:
Sumathi K, chaitra_udipi@yahoo.com
Aruna Rao K, arunaraomu@yahoo.com

Editor: Mohanad Alkhasawneh,mohanadf_73@yahoo.com

READING THE ARTICLE: You can read the article in portable document (.pdf) format (131927 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 1732 times since AUGUST 12, 2009.


Return to the InterStat Home Page.