Referential choice between common place nouns and place names for urban place types in property listings: A spatial analysis from the accessibility perspective

This study explores spatial association in the use of reference types for subway stations, shopping centers, parks, and health centers within the Persian language, specifically in Tehran, Iran. Using global and local join count statistics, we identify clustered patterns in the use of proper and common nouns. We employ a random forest model to predict referential choice based on geospatial and textual features, specifically for subway stations. Our prediction analysis demonstrates the feasibility of forecasting referential choice, with recall and precision values of 0.66 and 0.68, respectively, enhancing understanding of how language interacts with spatial contexts.