A real study-based modeling of stochastic behavior of traffic crash counts using penalized poisson-GzLM

Abstract

Data-based prediction models for vehicular crash counts are in high demand by transport and traffic authorities in Qatar. The road crash models based on in-depth data are important for developing efficient road safety analysis and auditing. The collection of such data is often expensive or even not possible. This work outlines the process through which the penalized maximum likelihood-based Poisson regression is applied to model the vehicular crash as a function of several categories of driving licenses issued in Qatar during the period 2007-2012. A real case study from Qatar is introduced and analyzed. 2019 IEOM society international.