CoPFun: an urban co-occurrence pattern mining scheme based on regional function discovery

Xiangjie Kong, Menglin Li, Jianxin Li, Kaiqi Tian, Xiping Hu, Feng Xia

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Analysis of mobile big data enables smart cities from aspects of traffic pattern, human mobility, air quality, and so on. Co-occurrence pattern in human mobility has been proposed in recent years and sparked high attentions of academia and industry. Co-occurrence pattern has shown enormous values in aspects of urban planning, business, and social applications, such as shopping mall promotion strategy making, and contagious disease spreading. What’s more, human mobility has strong relation with regional functions, because each urban region owns a major function to offer specialized services for city’s operations and such location-based services attract massive passenger flow, which is exactly the essence of urban human mobility pattern. Therefore, in this paper, we put forward a co-occurrence pattern mining scheme (CoPFun) based on regional function discovery utilizing various mobile data. First, we do traffic modeling to map trajectory data into population groups, which include temporal partition and map segmentation. Then we employ a frequent pattern mining algorithm to mine co-occurrence event data. Meanwhile, we exploit TF-IDF method to process POI data and LDA algorithm to process trajectory data to discover urban regional functions. We apply CoPFun to real mobile data to extract co-occurrence event data and compare it with OD data to analyze urban co-occurrence pattern from a perspective of regional functions. The experiment results verify the effectiveness of CoPFun.

Original languageEnglish
Pages (from-to)1029-1054
Number of pages26
JournalWorld Wide Web
Volume22
Issue number3
DOIs
Publication statusPublished - May 2019

Fingerprint Dive into the research topics of 'CoPFun: an urban co-occurrence pattern mining scheme based on regional function discovery'. Together they form a unique fingerprint.

Cite this