Bayesian Analysis of Zero-Inflated Count Data Using Beta-Lindley Distribution
Atheer Ismael Fadhel, Hossein Jabbari Khamnei · 2024
The process of analyzing data of large numbers assumes the presence of a unique and large set of challenges due to extra zeros and excessive dispersion processes. Therefore, data types that rely on the traditional counting principle attempt to capture the characteristics of this data appropriately and work on it bypassing these problems. Hence, the Bayesian principle was proposed, which seeks To model inflated zero count data based on the beta Lindley distribution, as this distribution is known for its flexibility and ability to deal with excess zeros and excess dispersion simultaneously, through the use of Markov chain Monte Carlo (MCMC) techniques for parameter estimation and model inference, as this paper seeks. The research aims to clarify the effectiveness of the proposed method by studying simulations and applying them to real-world data sets. The results of the Beta-Lindley distribution provide a robust and flexible framework based on the principle of modeling zero-inflated count data. This approach will provide accurate parameter estimates and reliable model predictions and will therefore have multiple uses in a range of different activities, including epidemiology, some environmental studies, and even within the economic sciences, where data are common.