Intelligent analysis method for traffic big data based on topology compression and feature dimension reduction

Abstract

The present application relates to the field of intelligent analysis of traffic big data, and provides an intelligent analysis method of traffic big data based on topology compression and feature dimension reduction, comprising: S1: obtaining time-series traffic state data of each road segment node in a target road network, and performing data cleaning and standardization processing, and dividing into a training set, a validation set and a test set; S2: based on the training set data, calculating the correlation between road segment nodes and constructing a road network topology structure containing a node layer and a hyperedge layer; wherein each node is uniquely attributed to a hyperedge; S3: based on the road network topology structure, using a set message passing mechanism to encode and compress node features, obtaining compressed features of nodes and hyperedges after dimension reduction; S4: fusing the compressed features of nodes and hyperedges, inputting a prediction model based on a graph attention mechanism, and predicting a future traffic congestion state of a target road segment. The present application solves the problems of high computing load, feature redundancy and insufficient prediction accuracy of large-scale traffic data.

Images