Weather risk contribution to traffic accidents types in Gulf Cooperation Council (GCC) countries

Traffic accidents had caused severe consequences for different road users worldwide. Several risk factors were attributed to traffic accidents occurrence. This study provided an overview of the weather effect on three selected traffic accident types; hitting pedestrians, accidents due to vehicle defects, and other types of crashes. The study was conducted in two stages. The first one included 9736 traffic accidents covering a two-year interval, from 2009 to 2010, in United Arab Emirates that were used to inspect the weather impact on the selected traffic accident types. Fog, rain, dust, and fine weather conditions were covered in relation to accident occurrence. A multinomial logit regression model was used to identify the correlations between the dependent and independent variables. A C5.0 tree was then applied to provide a performance comparison between the two models. The second stage was over three-year period from 2016 to 2019, which included more balanced data to validate the model. More classification methods were applied including C5.0, Random Forest, and K-nearest neighbors’ algorithm. The temporal variable (time in years) showed no significant relationship with traffic accident types. However, the weather conditions showed a significant impact on the types of traffic accidents. Fine weather conditions showed higher probabilities of pedestrian traffic accidents. Regarding model performance, C5.0 showed a higher prediction accuracy compared to the other applied models in the two stages.