Georeferencing Wikipedia documents using data from social media sources
- Submitting institution
-
Cardiff University / Prifysgol Caerdydd
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 95782277
- Type
- D - Journal article
- DOI
-
10.1145/2629685
- Title of journal
- ACM Transactions on Information Systems
- Article number
- 12
- First page
- -
- Volume
- 32
- Issue
- 3
- ISSN
- 1046-8188
- Open access status
- Out of scope for open access requirements
- Month of publication
- June
- Year of publication
- 2014
- URL
-
http://dx.doi.org/10.1145/2629685
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
A - Artificial intelligence and data analytics
- Citation count
- 9
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- While a considerable proportion of queries on the web are geographically specific (estimated about 20%), most documents that refer to geographical space are not accompanied by supporting geo-reference data that records location. This paper presents a novel, social-media based language modelling technique for geo-referencing that, unlike conventional methods, exploits all text in the document to infer location. A comprehensive set of experiments illustrates the benefits in locational accuracy of the presented approach. A literature review highlighted the method as an “innovative approach” (https://doi.org/10.1111/tgis.12212), with the best median error performance, and such language modelling approaches continue to gain momentum.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
