Design of Dynamic Traffic Information Prediction System Based on Multi-Source Information Fusion

With the rapid development and application of artificial intelligence technology, the field of traffic management has gradually entered the fast lane of intelligent transformation and upgrading. Facing the increasingly complex traffic network environment, it is urgent to use a series of new data information fusion methods to realize the effective prediction of dynamic traffic information, strengthen the management and scheduling of traffic flow, and improve the capacity of urban traffic. In this regard, this paper complete the application research of deep learning method in dynamic traffic information prediction, and propose and implement a network based on LSTM. At the same time, the author designs a dynamic traffic information forecasting system based on B/S architecture. It is convenient for users to obtain road traffic forecasting information in the future through visual interactive mode. The simulation test proves that the road travel time prediction model based on LSTM proposed in this paper is superior to the traditional SVM and FCM mining algorithms in prediction accuracy. The constructed system can also show the urban road traffic information in an intuitive and visual form. It greatly improves the analysis and application ability of multi-source dynamic traffic information, thus reducing the probability of traffic accidents and making contributions to the construction of intelligent transportation.