Extracting hierarchical boundaries of places from noisy geotagged user-generated content

Abstract

A place reflects the collective cognition of the geographical extent and semantics of a named spatial domain, acting as a vital reference to a particular space in daily discourse. Boundaries and toponyms are essential identifiers of places. Frameworks that are efficient in real-world boundary determination of cognitive places are still missing. The emergence of a large amount of geotagged user-generated content (geo-UGC) offers new opportunities to model the place boundaries from a more human-centric perspective. However, the broad geographical scales of places and the noise in geo-UGC data conflict with traditional approaches that only focus on places with similar spatial extents. In this paper, we advocate considering spatial hierarchy when determining place boundaries. We propose the Hierarchical Place Detector (HPD), a framework that composes noise detection, spatial hierarchy reconstruction, and boundary extraction, to rebuild the boundaries and spatial hierarchy of places from geo-UGC. The HPD is a state-of-the-art framework for determining and organizing place boundaries by the spatial hierarchy, thereby preserving morphology and geographical relationships among places. The hierarchical boundaries could be fundamental analytical units in various downstream applications, including spatial visualization, geographical information retrieval, navigation services, and spatial interaction modelling.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.