Parametric and Non-Parametric Analyses for Pedestrian Crash Severity Prediction in Great Britain

The study aims to investigate the factors that are associated with fatal and severe vehicle–pedestrian crashes in Great Britain by developing four parametric models five non-parametric tools predict crash severity. Even though have already been applied model pedestrian injury severity, a comparative analysis assess predictive power of such modeling techniques is limited. Hence, this contributes road safety literature comparing their capabilities identifying significant explanatory variables, performances terms F-measure, G-mean, area under curve. analyses were carried out using data refer occurred period 2016–2018. confirm advantages offering easy-to-interpret outputs understandable relations between dependent independent whereas exhibited higher classification accuracies, identified more provided insights into interdependencies among factors. results suggest combined use methods may effectively overcome limits each group methods, satisfactory prediction accuracies interpretation contributing serious crashes. In conclusion, several engineering, social, management countermeasures recommended.