Short-term passenger flow prediction for urban rail systems: A deep learning approach utilizing multi-source big data

Area of Science:

• Artificial Intelligence
• Transportation Science
• Data Science

Background:

• Intelligent and real-time management of urban rail systems relies on accurate short-term passenger flow prediction.
• Existing methods may not fully capture the complex spatial and temporal dynamics of passenger movement.

Purpose of the Study:

• To develop an enhanced spatial-temporal long short-term memory (ST-LSTM) model for forecasting subway passenger flow.
• To integrate multi-source big data and deep learning for improved prediction accuracy.

Main Methods:

• Developed an enhanced ST-LSTM model with three modules: temporal correlation learning, spatial correlation learning, and data fusion.
• Utilized multi-source big data and geographic information for spatial-temporal feature extraction.
• Evaluated model interpretability and performance.

Main Results:

• The enhanced ST-LSTM model effectively captures complex spatial-temporal correlations in passenger flow.
• The model demonstrated significantly superior performance compared to benchmark methods on real-world datasets from Nanjing and Chongqing.

Conclusions:

• The proposed ST-LSTM model offers a robust and accurate solution for predicting short-term subway passenger flow.
• This approach enhances the intelligent management capabilities of urban rail transit systems.