Short-term traffic flow prediction based on neural network projection pursuit coupling model

To address the strong randomness and nonlinear characteristics of short-term traffic flow, this paper proposes a short-term traffic volume and travel time prediction model based on neural network projection pursuit coupling (BPPP).The model combines the advantages of projection pursuit regression and BP neural networks and uses projection pursuit to process high-dimensional data. The model parameters are optimized using a real-number-coded genetic algorithm. Finally, experimental verification and model comparison are carried out using measured data from the Chongqing expressway. The experimental results demonstrate that the coupled model further reduces short-term traffic flow prediction error, with improvements in Error Count, Mean Absolute Error, Mean Absolute Percentage Error, and Root Mean Squared Error. This demonstrates the effectiveness of the proposed model.