Premium
Space–Time Sequential Similarity for Identifying Factors of Activity‐Travel Pattern Segmentation: A Measure of Sequence Alignment and Path Similarity
Author(s) -
Cho SungJin,
Janssens Davy,
Joh ChangHyeon,
Kim Hyunmyung,
Choi Keechoo,
Park Dongjoo
Publication year - 2019
Publication title -
geographical analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.773
H-Index - 65
eISSN - 1538-4632
pISSN - 0016-7363
DOI - 10.1111/gean.12186
Subject(s) - similarity (geometry) , pairwise comparison , segmentation , path (computing) , trajectory , computer science , cluster analysis , space (punctuation) , sequence (biology) , chaid , similarity measure , measure (data warehouse) , artificial intelligence , data mining , distance matrix , geography , pattern recognition (psychology) , algorithm , decision tree , image (mathematics) , physics , astronomy , biology , genetics , programming language , operating system
The article develops a new method that compares activity‐travel patterns in both terms of the sequential order of activities and the shape of activity‐travel trajectory in time and space. The similarity of the list of activities and their order between activity‐travel patterns are computed by a sequence alignment method. The shape of activity‐travel trajectory is compared between the patterns using a path similarity technique that captures the direction and speed of a movement from the current location and the duration of staying at each location. The comparison results, therefore capture how people move around in three‐dimensional space–time choreography that indicates how people conduct which activities in what order. A total of 1,000 individuals are sampled from the data of 2016 Household Travel Survey, South Korea. The data provide the information of individual activity‐travel behavior and personal characteristics. The suggested method computes the pairwise distance matrix, and Ward clustering algorithm segments the pattern groups of similar activity sequences and space–time trajectories. A CHAID analysis then associates personal and household characteristics with the pattern groups to identify important factors for the segmentation. The analysis provides a significant implication in both terms of research and practice in transportation.