A Bayesian analysis of accident data is used in the identification of hazardous locations. The Bayesian model used in the analysis is developed and discussed. Empirical comparisons of the results from the Bayesian analysis and from classical statistical analyses are also included. These comparisons suggest that there is an appreciable difference among the various identification techniques and that some classically based statistical techniques may be prone to err in the direction of false negatives.
Samenvatting