Identification model of urban road traffic safety risk sources based on big data

The identification of urban road traffic safety risk sources is an important link in ensuring transportation safety. In order to improve the identification accuracy of risk sources and reduce the rate of false positives, the paper proposes a new urban road traffic safety risk source identification model based on big data. Firstly, by combining radar and checkpoint systems, urban road traffic big data is collected, which mainly includes traffic flow data, average speed data, and occupancy data. Secondly, based on the collection results of transportation big data, extract the feature vectors of transportation big data. Finally, taking the extraction results of traffic big data feature vectors as input and the identification results of traffic safety risk sources as output, a convolutional neural network is used to construct an urban road traffic safety risk source identification model. The experimental results show that the average risk source identification accuracy of the proposed method reaches 95.56%, and the highest false alarm rate does not exceed 0.5%.