Prediction of Hangzhou Subway Station Passenger Flow based on Data Mining

Zhang, Pengfei (2020) Prediction of Hangzhou Subway Station Passenger Flow based on Data Mining. [Dissertation (University of Nottingham only)]

[img] PDF - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (6MB)

Abstract

An accurate passenger flow prediction is essential for subway station operators and passengers because it can reduce the congestion of subway stations, ensure passenger safety, and reduce passengers’ waiting time. The primary objective of this research is analysing the smart card data of Hangzhou subway stations and developing two prediction models for the passenger flow of subway stations in 15 minutes. The two models are linear regression model and the neural networks model. The testing and evaluation of these two models indicate that neural networks model has superior predictive accuracy than the linear regression model. During modelling, some researches are used to improve prediction performance. Firstly, this research explores the regularity of passenger flow in the subway station in different time granularity and use the Pearson correlation coefficient as the index of the regularity. The result indicates that the regularity of passenger flow in the subway station is better in the time granularity more than 15 minutes. Therefore, this research predicts the passenger flows in 15 minutes. Secondly, this study analyses different factors that affect passenger flow in the subway station. Thirdly, this research uses the K-means clustering algorithm to cluster the 80 subway stations into four station types with different passenger flow patterns. The station type is regarded as independent variable and inputs in two models. In addition, this research discusses the limitation of two models and proposes some improvements for the two models.

Item Type: Dissertation (University of Nottingham only)
Depositing User: Zhang, Pengfei
Date Deposited: 18 Apr 2023 13:55
Last Modified: 18 Apr 2023 13:55
URI: https://eprints.nottingham.ac.uk/id/eprint/62857

Actions (Archive Staff Only)

Edit View Edit View