Count Regression Models of Human Death by Road Traffic Accidents in Addis Abeba Ethiopia
No Thumbnail Available
Date
2015-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Abeba university
Abstract
Poisson regression model is commonly used to study the association between count outcome variable and covariates. This assumption does not hold true in real dataset due to the presence of overdispersion and excess zeros. While negative binomial regression model can deal with overdispersion, on the other hand, zero-inflated Poisson and zero-inflated negative binomial regression model can be used to handleexcesszeros in the observations. The objective of this study is to identify the most important traffic accident variables using count regression models and identify the best count data models in order to analyze road traffic accident datawhich is obtained from the Addis Ababa Traffic Control and Investigation Department (AATCID) daily basis recorded within 365 consecutive days from July 30, 2013 - July 29, 2014. Based on different model comparison, (using AIC and Vuong test) ZIP regression model provides more appropriate fit to the road traffic accident (the number of human death per RTA) data considered in this study. Month in quarter, road inclination, age of the drivers, vehicle type, ownershipof the vehicle, accident time, accident type and injured vehicle count were found to be statistically significant predictors of human deathat .a=0.05
Description
Keywords
Human Death by Road Traffic Accidents in Addis Abeba