Poisson Lindley process and its main properties (Q2322649)

The nonhomogeneous Poisson process, as a counting process for modeling random recurrent events, has several limitations as, e.g., the variance and the mean of the number \(N(t)\) of events in \([0,t]\) are equal, the process has independent increments. Hence, the aim of the article is to propose a modified counting process without these limitations and, in addition, is mathematically tractable. This is achieved by extending the Poisson Lindley process (PLP) to a counting process by using a mixing of (PLP) with certain additional terms. The properties of the modified process are discussed. Especially, it is shown that the modified process allows an explicit expression of the likelihood function.