Research on the Passenger Flow Distribution Mining of Urban Rail Transit Based on Clustering Algorithm: Taking Beijing as an Instance

Jiageng Ma; Bingxin Li; Yingqi Liu

doi:10.54097/2rfc5r29

Authors

Jiageng Ma
Bingxin Li
Yingqi Liu

DOI:

https://doi.org/10.54097/2rfc5r29

Keywords:

Passenger flow distribution, Urban rail transit, Clustering algorithm, Data mining.

Abstract

Based on the principal component analysis and k-means cluster analysis, the paper classified and analyzed the 310 stations of the Beijing subway. According to data such as the number of people entering and leaving each station of the Beijing Metro, the line to which the station belongs, and whether it is a transfer station by period. We selected a total of 52-time variables to participate in the PCA analysis and finally got two principal components and named them principal component 1 and principal component 2. Subsequently, we clustered the Principal Component 1 and Principal Component 2 scores of 310 sites and the latitude and longitude of the sites in k-means clustering. According to the elbow method and the contour coefficient, the final clustering results into five clusters are classified. The five clusters of stations have strong regional characteristics, and the stations of each type are more concentrated. The results of principal component analysis and cluster analysis have passed the test and have good convincing power. Finally, we conduct an in-depth analysis of the two categories with the most stations. Cluster 4 contains the largest number of stations, located in the core area of Beijing's urban area, and the passenger flow is significantly higher than other areas. Cluster 1 contains the second largest number of stations, mainly including Fengtai District. The region's economy is underdeveloped, the passenger flow is small, and there are peak and trough periods. This paper performs cluster analysis on high-dimensional data. The principal component analysis is used to reduce the dimension of high-dimensional data and retain the time and space dimensions required for clustering. The final result was in line with expectations. It is improved the shortcomings of k-means clustering for high-dimensional data. This article analyzes the daily passenger flow in Beijing, and the analysis results may be random. In the subsequent analysis, more data can be added to make the analysis more accurate.

Downloads

Download data is not yet available.

References

Wang Chuncai. Study on the Interaction Mechanism between Urban Traffic and Urban Space Development [D]. Beijing Jiao Tong University, 2007.

Zhao Hui, Yang Jun, Liu Changping. takes Beijing as an example [J].Economic geography, 2011,31 (09): 1445-1451.

Zhang Xiaodong, Li Shuang. World Urban Transportation System Comparison Study —— takes New York, London and Tokyo [J]. Beijing planning and construction, 2010 (06): 30-34.

Pan Tsunami, Ren Chunyang. Spatial coupling relationship between rail transit and urban public activity center system takes Shanghai as an example [J]. Urban Planning Journal, 2005 (04): 76-82.

Zhang Fina. Development trend of Urban rail transit in the context of big data [J]. The World of Digital Communications, 2020 (03): 149.

Xu Yuping, Qin Gong, Zhang Zheng.The Big Data Application Study of the Urban Rail Transit Survey [J]. Railway Transport and Economy, 2015,37 (04): 78-81.

Ran Jian. Development trend of Urban rail transit in the context of big data [J]. Low Carbon World, 2019,9 (08): 259-260.

Ma Limei, Dan, Gao Zhiyuan, Li Huajie.Big Data Technology and Its Industry Application: Concept Framework Research Based on Railway Field [J]. Journal of Beijing Jiao Tong University (Social Science Edition), 2019,18 (03): 58-67.

Jiang Ziwang. The Application of Big Data Technology in Urban Rail Transit Operation and Management [J]. Technological Innovation and Application, 2020 (05): 174-175.

Yu Chengcheng, Lv Hongxia. Research on Passenger Marketing Strategy Based on Big Data Railway Passenger Portrait [J]. Journal of Railways, 2020,42 (08): 23-28.

Huertas J A, Palacio A, Botero M, et al. Level of traffic stress-based classification: A clustering approach for Bogotá, Colombia[J]. Transportation Research Part D: Transport and Environment, 2020, 85: 102420.

Sfyridis A, Agnolucci P. Annual average daily traffic estimation in England and Wales: An application of clustering and regression modelling[J]. Journal of Transport Geography, 2020, 83: 102658.

Alisoltani N, Zargayouna M, Leclercq L. A sequential clustering method for the taxi-dispatching problem considering traffic dynamics[J]. IEEE Intelligent Transportation Systems Magazine, 2020, 12(4): 169-181.

Ryu U, Wang J, Pak U, et al. A clustering based traffic flow prediction method with dynamic spatiotemporal correlation analysis[J]. Transportation, 2021: 1-38.

Jin J, Wang W. Influential features PCA for high dimensional clustering[J]. The Annals of Statistics, 2016, 44(6): 2323-2359.

Dey S, Das S, Mallipeddi R. The Sparse MinMax k-Means Algorithm for High-Dimensional Clustering [C]// IJCAI. 2020: 2103-2110.

Dey S. Sparse min-max k-means algorithm for high-dimensional data clustering [J]. 2018.

Paul D, Chakraborty S, Li D, et al. Principal Ellipsoid Analysis (PEA): Efficient non-linear dimension reduction & clustering [J]. arXiv preprint arXiv: 2008.07110, 2020.