Volume : VIII, Issue : VIII, August - 2018

ZERO TRUNCATED COUNT MODELS WITH AN APPLICATION TO THE DENTAL CARIES DATA

Dr. S. B. Javali, Dr. C. M. Math

Abstract :

Count response data often exhibit departures from the assumptions of standard Poisson generalized linear models and ordinary least–squares estimation depends on the purpose of that estimation. The propensity of Decayed Missing Filled (DMF) index count data is to contain many zeros and to follow highly skewed distribution or overdispersed. Extra zeros however, violate the variance–mean relationship of the Poisson errors structure.The distribution of DMF counts had nearly 50% of zero values often are not observed.   In such case, the appropriate model for modelling positive DMF count data would be the models truncated at zero. This paper examines maximum likelihood regression estimators for DMF count data from truncated samples.  Estimators for the Zero Truncated Poisson and Negative Binomial regression models are illustrated and compared with standard Poisson regression and Negative Binomialregression models.  Truncated models can provide new insight to caries pattern in an examination of covariates on DMF positive counts.  It is anticipated that, Zero truncated models are becoming increasingly useful in recent epidemiological studies of dental caries data with positive counts as an outcome measure.

Keywords :

Article: Download PDF   DOI : 10.36106/ijar  

Cite This Article:

Dr. S. B. Javali, Dr. C. M. Math, ZERO TRUNCATED COUNT MODELS WITH AN APPLICATION TO THE DENTAL CARIES DATA, INDIAN JOURNAL OF APPLIED RESEARCH : Volume-8 | Issue-8 | August-2018

Number of Downloads : 839

References :