< Back to previous page

Publication

Developing a detection and interpretation tool for rumors regarding COVID-19 on Twitter

Book Contribution - Book Abstract Conference Contribution

Nowadays people use social media as one of their primary sources of information. Considering encountered misinformation can be detrimental to communications made by official instances, it is of major importance to detect potential rumors as soon as possible. The introduction of the term ‘infodemic’ by the World Health Organization highlights the magnitude and severity of the spread of misinformation during the COVID-19 pandemic. The aim of this study is to introduce a start-to-end rumor detection framework for COVID-19-related discussions on Twitter. Specifically, we present an approach for collecting Twitter data, building a suited detection model and gaining insights into the most important rumors. We investigate different traditional machine learning, deep learning and transformer models to determine which model is best suited for the classification of tweets into rumors and non-rumors in the context of COVID-19. Furthermore, we use a joint dimensionality reduction and clustering algorithm to cluster the COVID-19-related tweets into homogenous groups. Finally, we combine the results of the two aforementioned analyses to discover which clusters contain most rumors.
Book: EURO 2021, 31st European Conference on Operational Research, Abstracts
Pages: 142 - 142
Publication year:2021