Research on urban traffic congestion identification based on machine learning

In recent years, the rapid development of big data technology has facilitated the innovation of methods for the collection, analysis and application of large-scale data. The combination of trajectory data analysis and machine learning technology allows for the creation of an urban traffic congestion recognition system. In this work, we propose an urban traffic congestion recognition system based on Graph Neural Network (GNN), which combines the topology of the urban road network with the floating vehicle trajectory data for modeling. Firstly, the trajectory data is preprocessed and feature extracted, and it is converted into the form of nodes and edges to construct a dynamic road map. On this basis, GNN is used to learn the graph structure data to capture the complex spatial dependencies between roads, so as to achieve accurate classification and recognition of traffic conditions in different road sections. The model gradually optimizes the accuracy of traffic congestion recognition by iteratively updating node representations.