Crisis detection from Arabic tweets

Alaa Ali H Alharbi, Mark Lee

Research output: Chapter in Book/Report/Conference proceedingConference contribution

87 Downloads (Pure)

Abstract

Social media (SM) platforms such as Twitter offer a rich source of real-time information about crises from which useful information can be extracted to support situational awareness. The task of automatically identifying SM messages related to a specific event poses many challenges, including processing large volumes of short, noisy data in real time. This paper explored the problem of extracting crisis-related messages from Arabic Twitter data. We focused on high-risk floods as they are one of the main hazards in the Middle East. In this work, we presented a goldstandard Arabic Twitter corpus for four highrisk floods that occurred in 2018. Using the annotated dataset, we investigated the performance of different classical machine learning (ML) and deep neural network (DNN) classifiers. The results showed that deep learning is promising in identifying flood-related posts.
Original languageEnglish
Title of host publicationProceedings of the 3rd Workshop on Arabic Corpus Linguistics
PublisherAssociation for Computational Linguistics, ACL
Pages72-79
Number of pages8
ISBN (Electronic)978-1-950737-32-1
Publication statusPublished - 22 Jun 2019
EventThe 3rd Workshop on Arabic Corpus Linguistics (WACL-3) - Cardiff, United Kingdom
Duration: 22 Jul 201924 Jul 2019

Conference

ConferenceThe 3rd Workshop on Arabic Corpus Linguistics (WACL-3)
Abbreviated titleWACL-3
Country/TerritoryUnited Kingdom
CityCardiff
Period22/07/1924/07/19

Fingerprint

Dive into the research topics of 'Crisis detection from Arabic tweets'. Together they form a unique fingerprint.

Cite this