Capturing Public Concerns about Coronavirus Using Arabic Tweets: An NLP-Driven Approach
Source of Publication
Proceedings - 2020 IEEE/ACM 13th International Conference on Utility and Cloud Computing, UCC 2020
This In order to analyze the people reactions and opinions about Coronavirus (COVID-19), there is a need for computational framework, which leverages machine learning (ML) and natural language processing (NLP) techniques to identify COVID tweets and further categorize these in to disease specific feelings to address societal concerns related to Safety, Worriedness, and Irony of COVID. This is an ongoing study, and the purpose of this paper is to demonstrate the initial results of determining the relevancy of the tweets and what Arabic speaking people were tweeting about the three disease related feelings/emotions about COVID: Safety, Worry, and Irony. A combination of ML and NLP techniques are used for determining what Arabic speaking people are tweeting about COVID. A two-stage classifier system was built to find relevant tweets about COVID, and then the tweets were categorized into three categories. Results indicated that the number of tweets by males and females were similar. The classification performance was high for relevancy (F=0.85), categorization (F=0.79). Our study has demonstrated how categories of discussion on Twitter about an epidemic can be discovered so that officials can understand specific societal concerns related to the emotions and feelings related to the epidemic.
Institute of Electrical and Electronics Engineers Inc.
Communication | Computer Sciences
Cloud computing, Classification performance, Computational framework, NAtural language processing, Nlp techniques, Public concern, Societal concerns, Three categories, Two-stage classifiers, Natural language processing systems
Bahja, Mohammed; Hammad, Rawad; and Amin Kuhail, Mohammed, "Capturing Public Concerns about Coronavirus Using Arabic Tweets: An NLP-Driven Approach" (2020). All Works. 829.
Indexed in Scopus
Open Access Type
Bronze: This publication is openly available on the publisher’s website but without an open license