Abstract

Statistical models with countable data, where the degree of dispersion increases, are widely used by many practitioners in this field. Among these models are the Negative Binomial Regression Model, Poisson Regression Model, and Generalized Poisson Regression Model, among others. These models are particularly useful in cases of over dispersion found in certain real-world scenarios, such as (the number of traffic accidents, the number of complaints, the number of medical cases, and other data). This occurs due to the inequality between the variance of the dependent variable and it’s mean. This research focuses on the Generalized Poisson Regression Model, which is one of these models that can handle the degree of dispersion present in the Poisson model. Through conducting statistical analyses, we aim to obtain the best results that help reduce the number of traffic accidents.

Key words: Poisson regression, Generalize Poisson regression, Maximum Likelihood Method, AICC