K-neighboring on Multi-weighted Graphs for Passenger Count Prediction on Railway Networks

Abstract

In this paper, a method of multi-weighted graphs learning for passenger count prediction in railway networks is presented. Traffic prediction can provide significant insights for railway system optimization, urban planning, smart city development, etc. However, affected by the complexity of the railway networks as well as other factors, including spatial, temporal, and other external ones, traffic prediction on railway networks remains a critical task. To achieve high prediction performance of the models and discover the correlation between the models' performance and features, we propose six heterogenerous weight graphs, i.e., connection graph, distance graph, correlation graph and their fused graphs to fully construct the spatial and geometrical features of the railway network. Two representative graph neural networks (GNN), that is, the graph convolutional network (GCN) and graph attention network (GAT) were implemented for evaluation. The evaluation results demonstrate that the proposed GAT model learning on the correlation graph achieves the best performance, as it can reduce the metrics of mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error metrics (MAPE) on average by 19.7%, 6.9%, 27.9% respectively. Moreover, to investigate the trade-off on the graph scale and the prediction performance, we improved our model to search the optimal K value of neighbor nodes for graph partition for optimization. The K-neighboring method shows small scale graph also has the opportunity to achieve promising prediction result. The effectiveness of the proposal including multiple weight graphs as well as the optimal K-value for neighboring also were investigated. It can provide the interpretability of the traffic prediction tasks on the railway network.