City association pattern discovery: A flow perspective by using cultural semantic similarity of place name

Abstract

The study of cultural similarity between cities is of great significance for understanding the association of cities. However, there are still the following challenges in measuring and expressing cultural similarities: finding reliable data source to reveal urban cultural features and designing effective analysis and visualization model for cultural similarity. As a cultural symbol, place name can effectively reflect urban cultural characteristics. Therefore, this paper constructs a cultural eigenvector based on the semantics of place names. A measurement method is proposed to identify the cultural similarity between different cities and to explore and express similar patterns. In the form of OD flow, a cultural semantic similarity flow (CSSF) is defined to formally express cultural relations between cities. The cultural semantic similarity flow contains three patterns, namely one-to-one patterns, one-to-many patterns, and one-to-all patterns. These patterns explain the form of cultural similarity among cities from different perspectives, as follows: (a) One-to-one patterns may well represent the potential path of cultural diffusion; (b) one-to-many patterns reveal the regional cultural center to some extent; (c) one-to-all patterns show the spatial differentiation of culture on global scale. This general analysis and visualization method aim to accurately quantitatively analyze cultural similarities and their forms.