A New Regularly Varying Discrete Distribution Generated by Waring-Type Probability

Abstract

In this paper, based on the discretization method, we construct a new 2-parameter regularly varying discrete distribution generated by Waring-type probability (2-RDWP). Some useful plots are displayed for the model. From the mathematical point of view, to suggest 2-RDWP as a new discrete probability distribution in bioinformatics, some statistical facts such as unimodality, skewness to the right, upward/downward convexity, regular variation at infinity and asymptotically constant slowly varying component are established for the model. We provide the conditions of coincidence of solution for the system of likelihood equations with the maximum likelihood estimators for the unknown parameters. Simulation studies are performed using the Monte Carlo method and Nelder–Mead optimization algorithm to obtain maximum likelihood estimations of the unknown parameters. Asymptotic expansion of the probability function with two terms is considered, and then the moment’s existence of integer orders is investigated. Finally, a real count data set is used to show the applicability of the new model compared to other models in bioinformatics.