A LATENT STRUCTURE ANALYSIS OF A CITY CENTER USING SPATIAL AND CATEGORICAL CO-OCCURRENCE INFORMATION OF FACILITIES

Abstract

This paper aims at evaluating the facility accumulation of city center, focusing on spatial and categorical co-occurrence information of facilities. We expressed the co-occurrence information by relational data. We revealed the latent structure of the co-occurrence information by Extended Stochastic Block Model. We validated our approach by a case study in Osaka city center (around Namba station). We conclude that 1) the facility accumulation of the target area was classified by latent classes based on the co-occurrence information, 2) each latent class indicates its quantitative property, 3) the relationship between the latent classes correspond with their geographic distribution.