A fuzzy rough approach to analyze the significance of semantic levels for building tags in OpenStreetMap

In the realm of volunteered geographic information (VGI), the existence of comparable tags, attributes, and values across diverse categories of geographic objects gives rise to major categorization challenges such as conceptual overlap and indiscernibility. Enhancing the semantic data retrieval of VGI relies on the semantic quality of descriptive content annotated for tagging geographic objects. The main focus of this study is analyzing the descriptive content of OpenStreetMap to assess the significance of semantic levels. The proposed methodology relies on fuzzy rough set calculations to determine the degrees of dependency and significance of semantic levels. Three indicators, namely, the significance of semantic levels, decreasing the heterogeneity of attributes, and replicability were defined and assessed for a subset of building‐related tags. Analyzing building‐related tags in OpenStreetMap unveiled the higher significance for simple object, similarity, purpose, and function levels. The value of decreasing the heterogeneity of attributes was calculated at 63%, and the average replicability indicator of important attributes was doubled. Based on the results, the significance of semantic levels was deemed fit to enhance semantic homogeneity and replicability.