The application of predictive distribution estimation in multiple-inflated poisson models to ice hockey data

Article type: Research Article

Authors: Sadeghkhani, Abdolnasser^{a; *} | Ahmed, S. Ejaz^b

Affiliations: [a] Departamento de estadística, ITAM, CDMX, México | [b] Department of Mathematics and Statistics, Brock University, St. Catharines, ON, Canada

Correspondence: [*] Corresponding author: Abdolnasser Sadeghkhani, Departamento de estadistica, ITAM, CDMX, Mexico. E-mail: [email protected].

Abstract: In this paper, we introduce a multiple-inflated Poisson distribution that can handle count data with multiple inflated values. We explore a Bayes predictive distribution for future observation under Kullback Leibler loss function and a class of shrinkage priors, along with plug-in type pmf estimators. To illustrate how well the proposed pmf estimators perform, we provide both a simulated study and a real example analyzing a dataset of National Hockey League (NHL) shootout losses in 2017/18.

Keywords: Bayes predictive distribution estimation, count data, Kullback Leibler loss function, multiple-inflated poisson distribution, plug-in pmf estimator, shrinkage prior

DOI: 10.3233/MAS-200483

Journal: Model Assisted Statistics and Applications, vol. 15, no. 2, pp. 127-137, 2020