A method for assessing the credibility of volunteered geographic information in case of flood crisis

Abstract

In this paper, we developed a score-based credibility assessment model which assesses the credibility of text and map-based VGI. We proposed parameters for assessing the credibility of VGI such as distance, hazard, and clustering. Then, the model was created based on the above-mentioned parameters, and the credibility scores of reports were computed. Furthermore, we categorized the reports into high-credible and low-credible categories, and accuracy of the model was determined by comparing the model's output with the real situation. To assess the credibility of VGI, we made use of reports from the Ushahidi project in Brisbane, Australia, which was hit by a severe flood in 2013 as our first dataset. Besides the aforementioned parameters, we examined the distance between the location of the user and location of the incident as a parameter for the credibility assessment of VGI. For doing so, in a case study, we developed a VGI-based website to collect data for the analysis. Results show that our model which was created by spatial clustering and flood hazard parameters can categorize the credibility classes with 92.6% accuracy. In addition, in the second study area, the credibility model using the proximity of the volunteer to the incident site parameter can assess the credibility classes of VGI with 90% accuracy.

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 — 2 citations exist upstream.

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