H. Sun, S. Cabras
The daily metro station occupancy is an important indicator to ensure the safe and smooth operation of the metro.
An accurate forecast is essential for staff work arrangements and handling of emergencies. This work uses a hierarchical Bayesian model to forecast metro passenger flow at Beijing Metro stations. This approach is different from general machine learning prediction algorithms providing a reliable measure of uncertainty about passenger flow. The model includes spatial random effects for stations and the days of the week. Essentially, it estimates a spatial-temporal Poisson process for areal counts. This model provides an accurate result on daily passenger flow, which helps Beijing metro improve and optimize passenger organization.
Keywords: Bayesian model,passenger flow,spatial-temporal,metro,forecast
Scheduled
Bayesian methods
June 7, 2022 4:50 PM
Audiovisual room