Inferential properties with a novel two parameter Poisson generalized Lindley distribution with regression and application to INAR(1) process.

Based on the well-known Poisson (P) distribution and the new generalized Lindley distribution (NGLD) developed by using gamma ( α , θ) and gamma ( α − 1 , θ) distributions, a new compound two-parameter Poisson generalized Lindley (TPPGL) distribution is proposed in this paper and thereon systematically explores the mathematical properties. Closed form expressions are assembled for such properties including the probability generating function, moments, skewness, kurtosis, etc. The likelihood-based method is used for estimating the parameters followed by a broad Monte Carlo simulation study. To further motivate the proposed model, a count regression model and a first order integer valued autoregressive process are constructed based on the novel TPPGL distribution. The empirical importance of the proposed models is confirmed through application to four real datasets. [ABSTRACT FROM AUTHOR]