At present, the detection of urban community structures is mainly based on existing administrative divisions, and is performed using qualitative methods. The lack of quantitative methods makes it difficult to judge the rationality of urban community divisions. In this study, we used complex network association mining methods to detect a city community structure by using the Origin-Destinations (OD) at traffic analysis zone (TAZ) level, and successively assigned all the TAZs into different communities. Based on the community results, we calculated the community core degree of each TAZ within every community, and then calculated the Traffic Core Degree and Location Core Degree indicators of the community based on OD passenger flow and spatial location relationship between communities. Finally, we analyzed the correlation among three indicators to ensure the rationality of the community structure. We used the city of Zhengzhou in 2016 as an example case study. For Zhengzhou, we detected a total of six communities. We found a relatively low correlation between Traffic Core Degree and Location Core Degree. Within each group, the correlation between community core degree and Traffic Core Degree was higher than that between community core degree and Location Core Degree, indicating that the urban community structure is more reasonably based on traffic characteristics. The development of a quantitative approach for determining reasonable city community structures has important implications for transportation planning and industrial layout.
Robust image hashing with compressed sensing and ordinal measures
