Publication

Abstract : The modelling of count data is extensively used in many fields of research. There is handful of zero-inflated probability models available in literature. Among these models, zero inflated Poisson distribution is one of the widely used models for modelling data with excess number of zeros. In all the zero-inflated models, one can have parameter called zero-inflation parameter which is in addition to the number of parameters in underlying distribution. The estimation of the zero-inflation parameter of the zeroinflated Poisson (ZIP) models by MLE do not have an explicit expression and solved iteratively by using modern computing techniques. In this paper, we proposed a probability based inflation estimator (PBIE) for making inferences about the inflation parameter of the ZIP model and also studied the performance of the proposed estimator for simulated data.