Research on Classification and Prediction of General Traffic Accidents on National and Provincial Highways Based on BP Neural Networks

With the continuous increase in the number of motor vehicles in China, the problem of traffic safety is becoming more and more prominent, and traffic accidents occur frequently, in order to understand the probability of accidents and possible impacts, this paper proposes a classification prediction method for general traffic accidents on national and provincial roads based on BP neural network. The data of traffic accidents on national and provincial roads from 2018 to 2020 are classified by the characteristics of the influencing factors, and the data are divided into a test machine and a validation set to train the BP neural network model, and the neural network models with one hidden layer and two hidden layers are evaluated by the confusion matrix, and the results show that the BP neural network model with one hidden layer is shorter in time consumption, and the difference in the accuracy of the model is not very large, so the BP neural network model with one hidden layer is more accurate. layer BP neural network model is more suitable for the classification prediction of general traffic accidents on national and provincial highways. In this paper, the BP neural network-based classification prediction method of general traffic accidents on national and provincial highways is of great significance to reduce the occurrence of accidents and the risk of accidents.