Please use this identifier to cite or link to this item:
Title: Investigating the relationship between earthquakes and online news
Authors: Camilleri, Stephen
Azzopardi, Joel
Agius, Matthew R.
Keywords: Information retrieval
Big data
Text data mining
Information visualization
Earthquake hazard analysis
Geological mapping
Issue Date: 2019
Publisher: Institute of Electrical and Electronics Engineers
Citation: Camilleri, S., Azzopardi, J., & Agius, M. R. (2019, June). Investigating the relationship between earthquakes and online news. 2019 IEEE Second International Conference on Artificial Intelligence and Knowledge Engineering (AIKE), 203-210.
Abstract: News agencies have the leading role of deciding what coverage is given on earthquakes. This study exploits the use of text mining tools to automatically map in near real-time earthquake events with newspaper articles. An application was developed to automatically download multilingual news articles published by 23 leading news agencies in real-time from six continents around the world. Articles discussing earthquakes are then identified; clustered using TF-IDF and cosine similarity; and processed for information extraction. Earthquake-related features that are mined include the magnitude, location and timestamp from the corresponding news reports, as well as the number of casualties, injured and structural damage caused. Cluster information extraction is also carried out for crossreferencing purposes. Each cluster is then mapped against earthquake events, aggregated from seismic readings stored by United States Geological Survey (USGS). The results are visualised on two dashboards and complemented with a series of tests evaluating the accuracy of the extraction of information and clustering of data, followed by conclusions and recommendations for future work in this domain area. This tool paves the way for a more in-depth analysis on the possible temporal correlation (how long the disaster and its effects remain discussed in the news) and spatial correlation (geographical distance between news sources reporting the event and the location of the disaster itself) that may exist between newspaper articles and earthquakes
ISSN: 9781728114880/19
Appears in Collections:Scholarly Works - FacICTAI

Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.