Knowledge graph construction for short-term traffic flow prediction at urban intersections in rainy and snowy weather

In rainy and snowy weather, traffic flow is influenced by factors such as visibility, slippery road surface, and driving habits, resulting in significant errors in short-term traffic flow prediction. To address this issue, research was conducted and a knowledge graph was constructed for predicting short-term traffic flow at urban intersections in rainy and snowy weather. Firstly, under the influence of rainy and snowy weather, the characteristic parameters of short-term urban traffic flow were calculated. Subsequently, a knowledge graph for traffic flow prediction was constructed through case analysis, ontology and entity layer construction, and storage and display of knowledge graphs. Finally, based on the constructed knowledge graph and short-term traffic flow prediction results, a short-term traffic flow prediction model was constructed using a relationship graph convolutional neural network. The experimental results show that the prediction method proposed in this paper significantly reduces prediction errors, with the lowest RMSE value being 1.045.