Application of short text topic modelling techniques to Greta Thunberg discussion on Twitter

Dingemans, Sean (2020) Application of short text topic modelling techniques to Greta Thunberg discussion on Twitter. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration manual) Download (12MB) | Preview

Abstract

Twitter has become an essential medium for probing differing views on issues within society. One such issue is climate change, which is of interest to social scientists. Topic modelling is a way to discover such discussion and viewpoints. Latent Dirichlet Allocation (LDA) is the most widely used topic modelling technique. However, there are limitations for LDA with texts typically less than fifty words, where it suffers from poor characterisation. Consequently, six novel topic modelling algorithms were evaluated against LDA to assess their performance. Topics around climate change activist Greta Thunberg were to be examined as a secondary objective. The collected data comprised of Tweets centring on her United Nations Climate Action speech. Topic allocations for the topic modelling algorithms were evaluated with a novel combination of classification recall and coherence scores. The algorithms Word Network Topic Model (WNTM) and Biterm Topic Model (BTM) were found to have the best overall performance. These novel algorithms will be of great interest to social scientists and marketing companies who would like to probe the discussions on Twitter better. Discussion around Greta Thunberg was found to be polarised.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software T Technology > T Technology (General) > Information Technology > Computer software
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Dan English
Date Deposited:	18 Jan 2021 14:00
Last Modified:	18 Jan 2021 14:00
URI:	https://norma.ncirl.ie/id/eprint/4366

Actions (login required)

View Item