Abstract:
Road traffic accidents are one of the leading causes of injuries and death in both developed and
developing countries. According to WHO, 1.35 million people die each year as a result of road
traffic accidents globally. Ethiopia is one of the developing countries and at least 114 people die
for every 10,000 vehicle accidents annually. Moreover, road traffic accident the capital city, Addis
Ababa resulting in thousands of physical injuries and costing the economy in millions of dollars.
Hence, time series analysis related to the road traffic accident has a very important place in
revealing the future trends of the accident and decision making process. Therefore, this study
focuses on statistical analysis of road traffic accident using Seasonal Autoregressive Integrated
Moving Average (SARIMA) and Self-Excited Threshold Autoregressive (SETAR) time series
models. Data were obtained from Addis Ababa Traffic Police Commission and temporally
aggregated from January 2004 to December 2018 for analysis purpose. Data analyses were
performed using R and S-plus statistical software. The estimated trend component of RTA showed
a rising trend from 2010 to 2016 G.C .Furthermore, road traffic accident most frequently occurs
during the rainy seasons (June, July and August) of Ethiopia. The two regime SETAR model was
adopted to accommodate non linearity and linear SARIMA model was fitted as a benchmark for
comparative analysis. The model was nominated from SARIMA and SETAR models based on the
selection criteria and model comparison was made between the selected models. Nonlinear
SETAR(2,8,8) outperformed forecast than linear SARIMA(1,1,1)(1,1,2) 12 model for road traffic
accident of Addis Ababa. The out sample forecasted value indicates that, road traffic accident has
an increasing trend over the forecasted period.
