A new, non-canonical Poisson regression model for the prediction of crashes on low-volume rural roads

The purpose of the study was to compare the prediction power of a simplified non-canonical Poisson crash-prediction model to other model types. The model, fitted to serious and fatal crash data from 86 two-lane low-volume rural highway segments, showed a good fit, which was not significantly different from that of a negative binomial model. The application of the present model uses the linear form of the non-canonical Poisson model. Hence the simplification of the model versus other models results from the finding that the expected number of crashes per 1 km is directly proportional to the daily volume, unlike logarithmic functions in other models. In the non-canonical model, it is necessary to estimate only one parameter, whereas estimations of more parameters are needed in the negative binomial model.

Figure optionsDownload as PowerPoint slideHighlights
► Distribution of number of fatal and serious crashes on low volume roads is Poisson.
► The average number is a product of a parameter beta, section length and ADT.
► The parameter beta depends on drivers' behavior and infrastructure characteristics.
► The parameter beta is easily estimated from field data of lengths and volumes.
► The model allows to identify easily overly dangerous highways that need remedy.