ABSTRACT Beijing is famous for the large number of people and the significant phenomenon of job-housing mismatch, and residents here have specific mode of travel, especially in the Beijing Urban Rail Transit (BURT). Multi-dimensional Frequent Pattern Mining (MFPM) can efficiently mine the frequent patterns of residents' travel behavior, find out the association rules of trips in BURT, and explore the underling mechanism. We obtained travel information from the transit card data of BURT in 2015 and chose data in spatial, temporal and line dimensions as the attribute sets. It’s found that there is some land use related “station group” such as “Xierqi Group” and “Guomao Group” and several strongly associated "commuting OD pairs" such as {Wangjing, Maquanying} and {Changyang, Fengtai}. The research will help unraveling the travel regularities of riders in BURT and assisting operation management. Keywords: association rule; Multi-dimensional Frequent Pattern Mining; smart-card data; trip; urban rail transit

CICTP