关键词: 城市交通, 车站精细化分类, 无监督聚类, 地铁车站, 土地开发

Abstract: This paper studies the fine classification of subway stations. In this paper, we use a commuter trip identification method based on trip chain analysis to identify the commuter passenger flow, then analyze the jobhousing characteristics of stations with the inbound and outbound passenger flow volume in the morning and evening peak. The point of interest (POI) data captured from the open-source platform of the Baidu map is classified according to the land use function to explore the built-up environment characteristics around the stations. Combined with the two types of characteristics, a subway station classification model based on the unsupervised learning k-means++ method is established, and 307 stations of Beijing subway are divided into 7 categories. According to the characteristics of passenger flow and surrounding land use, the stations are classified as the typical residential station with welldeveloped supporting facilities, the typical residential station with commercial development potential, the residential station configuring certain jobs, the highly developed typical working station, the working station combined with work and housing, the tourist and leisure station and the outer suburban station to be developed. The classification results are highly consistent with the actual situation, which verifies the effectiveness of the model, and can provide the basis for city planning and land development around the station.

Key words: urban traffic, station fine classification, unsupervised clustering, subway station, land development