Comparison of spatial regression models with Road Traffic Accidents Data

Road traffic accidents (RTA) cause severe problems for the societies in developed and developing countries and result in loss of lives and highly burden cost. Investigators have applied a wide variety of methodological techniques over the years in order to gain more understanding in this discipline. Spatial models, an elegant technique to model spatial data, has been applied to capture localisation effects and influencing factors on RTA. This study aims to compare the spatial lag model (SLM) and the spatial error model (SEM) with the Ordinary least square (OLS) model for the road traffic accident data in the Sultanate of Oman. To compare the performance and accuracy of the OLS, SLM and SEM models, the log-likelihood, Akaike information criterion (AIC), and Bayesian information criteria (BIC) have been used. The SLM model outperforms SEM and OLS models for the data and one of the most significant findings to emerge from this study as the SLM model provides the best values of log-likelihood, AIC and BIC comparing to the values from OLS and SEM models