Metro Station Classification by Service Function Based on AFC Data and RF Method

Abstract

Abstract:

The spatial distribution of the type of passengers rail stations serving is an indicator of urban structure. This paper analyzes the spatial and temporal evolution of the type of passenger flow on the whole network station service by using random forest model (RF) and automatic fare collection (AFC) data. For the traditional RF model relying too much on subjective experience in selecting training sets, this paper firstly uses traditional RF model to measure the degree of similarity between stations, and then adopts the PAM method to cluster. Clustering results show that the unsupervised RF method has better accuracy. Finally, based on a large amount of AFC data and unsupervised RF method, we reveal the change of metro station attribute from 2014 to 2017 in Beijing. The analysis shows that the city’s structure of job and house location in Beijing has largely remained unchanged on a large spatial scale in the past four years, but the functional areas within the city are undergoing a slow variation, and the area where stations serving both commuting and living passengers are hot spots in the evolution of urban structure. The research results of this paper can provide a new perspective for the understanding of the interaction between urban rail transit and urban structure.

Key words: urban traffic, unsupervised random forest, station classification, similarity measure, evolution

摘要：

城轨线网上服务不同类型客流车站的分布能反映城市结构特征，本文基于多年的城轨AFC数据，应用随机森林模型(RF)，分析了全网车站服务客流类型的时空演变.对于传统RF在选择训练集时过于依靠主观经验的问题，本文首先利用传统随机森林度量车站之间的相似性，再采用Partitioning Around Medoid(PAM)方法对度量结果进行聚类，结果表明无监督随机森林方法拥有更好的准确性；最后采用无监督随机森林对北京市2014—2017年的车站服务客流属性的时空变化情况进行分析，结果显示，北京市近4年来城市职住结构在大的空间尺度上基本保持不变，但城市内部各功能区正经历缓慢变化，而服务于居住与工作混合类客流的车站所处区域将是今后城市结构演变中热点区域.本文的研究结果可以为城市轨道交通与城市结构互动关系的认识提供新视角.

关键词: 城市交通, 无监督随机森林, 车站分类, 相似性度量, 演变

CLC Number:

U231

WANG Zi-jia, LIU Hai-xu, TAKU Fujiyama. Metro Station Classification by Service Function Based on AFC Data and RF Method[J]. Journal of Transportation Systems Engineering and Information Technology, 2018, 18(4): 224-230.

王子甲，刘海旭，TAKU Fujiyama. 基于AFC数据和RF模型的城轨车站服务功能分类[J]. 交通运输系统工程与信息, 2018, 18(4): 224-230.

Add to citation manager EndNote|Ris|BibTeX

URL: http://www.tseit.org.cn/EN/abstract/abstract19684.shtml

http://www.tseit.org.cn/EN/Y2018/V18/I4/224

References

[1] PELLETIER M P, TRÉPANIER M, MORENCY C. Smart card data use in public transit: A literature review[J]. Transportation Research Part C: Emerging Technologies, 2011, 19(4): 557-568.

[2] LEE S, HICKMAN M D. Travel pattern analysis using smart card data of regular users[R]. Transportation Research Board Meeting, 2011.

[3] BRIAND A S, CÔME E, TRÉPANIER M, et al. Analyzing year-to-year changes in public transport passenger behaviour using smart card data[J]. Transportation Research Part C: Emerging Technologies, 2017(79): 274-289.

[4] CEAPA I, SMITH C, CAPRA L. Avoiding the crowds: understanding tube station congestion patterns from trip data[C]//Proceedings of the ACM SIGKDD International Workshop on Urban Computing. ACM, 2012: 134-141.

[5] UTSUNOMIYA M, ATTANUCCI J, WILSON N. Potential uses of transit smart card registration and transaction data to improve transit planning[J]. Transportation Research Record: Journal of the Transportation Research Board, 2006 (1971): 119-126.

[6] BREIMAN L. Random forests[J]. Machine learning, 2001, 45(1): 5-32.

[7] 方前程, 商丽, 商拥辉, 等. 爆破振动诱发民房结构损伤识别的随机森林模型[J]. 爆炸与冲击, 2017, 37(6): 939-945. [FANG Q C, SHANG L, SHANG Y H, et al. Random forest model for identification of residential structure damage induced by blast vibration[J]. Explosion and Shock Waves, 2017, 37(6): 939-945.]

[8] 钱超, 陈建勋, 罗彦斌, 等. 基于随机森林的公路隧道运营缺失数据插补方法[J]. 交通运输系统工程与信息, 2016, 16(3): 81-87. [QIAN C, CHEN J X, LUO Y B, et al. Random forest based operational missing data imputation for highway tunnel[J]. Journal of Transportation Systems Engineering and Information Technology, 2016, 16(3): 81-87.]

[9] REBOLLO J J, BALAKRISHNAN H. Characterization and prediction of air traffic delays[J]. Transportation Research Part C: Emerging Technologies, 2014(44): 231-241.

[10] 岳真宏, 陈峰, 王子甲, 等. 基于刷卡数据和高斯混合聚类的地铁车站分类[J]. 都市快轨交通, 2017, 30(2): 48-51, 107. [YUE Z H, CHEN F, WANG Z J, et al. Classifications of metro stations by clustering smart card data using the gaussian mixture model[J]. Urban Rapid Rail Transit, 2017, 30(2): 48-51, 107.]

[11] 北京市交通委. 第五次北京城市交通综合调查报告 [R]. 北京: 北京市交通委, 2016. [Beijing Municipal Commission of Transport. The fifth Beijing urban transport comprehensive survey report[R]. Beijing: Beijing Municipal Commission of Transport, 2016.]

[12] 周绮凤, 洪文财, 杨帆, 等. 基于随机森林相似度矩阵差异性的特征选择[J]. 华中科技大学学报(自然科学版), 2010, 38(4): 58-61. [ZHOU Q F, HONG W C, YANG F, et al. Feature selection of random forestbased proximity matrix difference[J]. Journal of Huazhong University of Science and Technology(Natural Science Edition), 2010, 38(4): 58-61.]

[13] BREIMAN L, CUTLER A. Random forests. 2004[J]. (2004-06-15) [2018-03-01]. https://www.stat.berkeley. edu/~breiman/RandomForests/cc_home.htm#unsup.

[14] SHI T, HORVATH S. Unsupervised learning with random forest predictors[J]. Journal of Computational and Graphical Statistics, 2006, 15(1): 118-138.

[15] VESANTO J, ALHONIEMI E. Clustering of the selforganizing map[J]. IEEE Transactions on Neural Networks, 2000, 11(3): 586-600.