Mapping geospatial events based on extracted spatial information from web documents

Abstract

Web documents such as news articles, social feeds, and blogs provide an abundant and readily available data source of spatial information relating to dynamic events such as wildfires, storms, and chemical spills. Research in the fields of geographic information retrieval and natural language processing use methods to extract place-names from web documents that can be used to geocode these events. However much of the spatial information in these articles are difficult to use because of the inherent vagueness of natural language. This thesis aims to develop methods to handle the vaguness of representing natural language descriptions of events by integrating precise spatial information (landmarks and geographic coordinates) with imprecise spatial information to provide a map-based visualization of the likely spatial extent and location of web document events.

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 — 3 citations exist upstream.

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