Cited by EF-former for Short-Term Passenger Flow Prediction During Large-Scale Events in Urban Rail Transit Systems

A Noise-Robust Approach Using Dynamic Graph Neural Networks for Bus Passenger Flow Prediction DOI

Xinyi Zhou, Siyu Sun, Nizar Bouguila

et al.

Published: Jan. 1, 2025

Short-term bus passenger flow prediction is essential for real-time operational optimization and efficient management of public transportation systems. This paper proposes a Noise-Cleaning Dynamic Graph Neural Network Liquid framework, which addresses noise caused by data anomalies such as sensor errors unexpected events or irregularities. The model leverages dynamic graph neural networks to capture spatial-temporal dependencies, integrating snapshots feature expansion connectivity changes. Additionally, liquid enhance temporal sequence adaptability, while noise-cleaning module mitigates variability using K-NN imputation Gaussian Mixture Models. validated real-world from the Ames system, incorporating route cancellations, holiday effects, weather variability. Compared baseline models, including CNN, GRU, LSTM, NC-DGNN-LNN framework demonstrates superior performance in accuracy, robustness, computational efficiency. Results suggest that proposed approach provides scalable accurate solutions improving prediction. code available at: https://github.com/XinyiZhou0318/A-Noise-Robust-Approach-Using-Dynamic-Graph-Neural-Networks-for-Bus-Passenger-Flow-Prediction.

Language: Английский

Citations

Short-Term Passenger Flow Prediction of Urban Rail Transit on Bayesian Optimization- Bidirectional Long Short-Term Memory with Causal Temporal Pattern Attention DOI