Short-time Prediction of Urban Rail Transit Passenger Flow

Abstract

Accurate prediction of short-term passenger flow in urban rail transit systems plays a crucial role in optimizing operations and enhancing passenger experience. This study presents a scientific approach to predict subway passenger flow by analyzing characteristic patterns, identifying key factors influencing passenger flow changes, and leveraging relevant data sources. The multi-source data used in this study are described and pre-processed to capture the spatial, temporal, and other factors that contribute to subway passenger flow distribution. Utilizing the extracted features as inputs, an improved Long Short-Term Memory (LSTM) method is employed for short-term passenger flow prediction. The performance of the improved LSTM method is compared and analyzed against traditional methods. The results demonstrate that the proposed approach outperforms traditional methods in terms of prediction accuracy for the same prediction target. Furthermore, the fusion of multi-source data and the inclusion of external factors significantly enhance the prediction accuracy. This research highlights the importance of considering various factors and data sources when forecasting short-term passenger flow in urban rail transit systems. By employing an improved LSTM method and integrating multiple data dimensions, the proposed approach offers superior prediction accuracy compared to traditional methods. The findings contribute to the development of efficient and reliable prediction models for optimizing urban rail transit operations and improving passenger services.

Keywords

• data mining
• LSTM
• prediction of short-time passenger flow
• rail transit