The Application of Generalized Linear Regression Models Cope with the Types of Vehicles in Road Accidents

Abstract :

In the modern era, Road accident prediction models are priceless tools that have valuable applications in road
accidents protective analysis. This paper focuses on the Generalized Linear Regression (GLR) modeling on the number of people died, injured with primarily involved vehicles by road accidents for the years 2001-2015 in Tiruchirappalli District and the exhaustive analysis of the data using two statistical techniques such as Poisson regression and Negative Binomial regression to fit a model to the data. To propose improvement measures to prevent road accidents and to derive a model for accident parameters. This paper suggests procedures for developing prudent models, i.e. models that are not overfitted, and best-fit models. The respective models were used to identify the vehicles that caused more people died and injured.

Author Name : Philip Arokiadoss. A, Muthu. C & Arulanandu. U

DOI: https://doi.org/10.5281/zenodo.345433

Keywords: GLR, Negative Binomial Regression Models, Number of people died, Number of People injured, Poisson Regression Model and Vehicles.