Bayesian Inference for Analyzing Sports Data by using Bivariate Distribution Models

Abstract

It is a common practice to use models based on the bivariate distributions for modeling sports data. These models are used to analyze discrete count data with two dependent variables in the data. In this dissertation we use Markov chain and Monte Carlo techniques to implement on simulated data and we perform real data analysis to demonstrate model fitting performances of our proposed models. We use Poisson regression (BP) and diagonally inflated bivariate Poisson regression (DIBP) models to analyze soccer data and we analyze baseball and basket ball data by using different types of bivariate binomial models (BVB).