Extracting Metro Passengers' Route Choice via AFC Data Utilizing Gaussian Mixture Clustering

Abstract

Metro system plays an essential role in urban public transportation. Studying and understanding how passengers make their path decision are of significance for metro system management. Thanks to the state-of-art of the Auto Fare Collection (AFC) system, every individual trip information is recorded. However this system only provides information of origin and destination spatial-temporal information, the specific route that chosen by every passenger are not available. In this paper, a passengers' routing behavior extracting method, which combines the analytical method and the data based method, is proposed. This method firstly models the travel time to obtain the effective path set based on some rules. By applying the Gaussian Mixture Model (GMM) clustering, the AFC date between an OD pair is assigned to every effective path such that the probability of being chosen for every effective path can be obtained. Our method can achieve a relatively high accuracy with low computing cost when comparing with other methods. Our method is applied to analyze the network of Beijing metro, simulation results demonstrate that the method is able to process large-scale AFC data and obtain the passengers' path choice pattern.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.