Scatter-GNN: A Scatter Graph Neural Network for Prediction of High-Speed Railway Station—A Case Study of Yinchuan–Chongqing HSR

The Yinchuan–Chongqing high-speed railway (HSR) is one of the “ten vertical and ten horizontal” comprehensive transportation channels in National 13th Five-Year Plan for Mid- Long-Term Railway Network. However, choice node stations on this line controversial. In paper, problem station selection transformed into a classification under edge graph structure complex networks, Scatter-GNN model proposed to predict stations. article first uses Node2vec algorithm perform biased random walk network generate vector representation each station. Secondly, an adaptive method proposed, which derives critical value through pinching rule, then effectively identifies network. Next, calculation Hadamard product used represent potential neighbors sites, attention mechanism link between all their corresponding sites. After prediction, final obtained, it input GNN classifier together with label complete prediction. Experiments show that: Baoji Hanzhong are more likely become north–south trunk line. optimizes site strategy by calculating connection probabilities two or candidate routes comparing results. This may reduce manual costs ease geographic evaluation burdens. paper can be as auxiliary traditional route planning scheme, new way thinking study such problems future.