Cluster Analysis of Trip Purpose Based on Residents' Travel Characteristic

In the era of big data and artificial intelligence, individual generated traffic data has gradually become the key data source for extracting personal behavior information and travel purpose prediction. Based on the Person Trip survey of residents' travel in Tokyo, this paper uses Spearman correlation coefficient to explore the correlation between residents' travel attribute patterns. In the situation of unsupervised learning, K-means clustering algorithm is used to analyze the differences of travel purpose caused by various attributes of residents' travel. This paper explores the possible reasons for the low accuracy of travel purpose prediction by supervised machine learning methods. Through the visual display of clustering results, it puts forward guiding suggestions for the method system of obtaining residents' travel patterns based on traffic big data, which is of great significance to the development of new technologies and academic theories in the field of urban transportation.