lnu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Machine Learning for Social Sciences: Stance Classification of User Messages on a Migrant-Critical Discussion Forum
Linnaeus University, Faculty of Social Sciences, Department of Social Studies. (DISA;CSS)ORCID iD: 0000-0001-9938-2675
Linnaeus University, Faculty of Technology, Department of computer science and media technology (CM). Linköping University, Sweden. (ISOVIS;DISA-DH)ORCID iD: 0000-0002-1907-7820
2021 (English)In: Proceedings of the 2021 Swedish Workshop on Data Science (SweDS) / [ed] Rafael M. Martins, Morgan Ericsson, Danny Weyns, Kostiantyn Kucher, IEEE, 2021, p. 1-8Conference paper, Published paper (Refereed)
Abstract [en]

In this paper, we present our methodology for supervised stance classification of sparse and imbalanced social media data. We test our framework on a manually labeled dataset of 5700 messages about immigration in the Swedish language posted on the Flashback forum, a controversial online discussion platform. Our proposed approach currently achieves a macro- averaged F1-score of 0.72 for test data on a two-class problem compared against 0.27 for a baseline four-class model. Since effective classification of imbalanced and sparse textual data in under-resourced languages presents certain methodological challenges, our study contributes to a discussion on the best pathways to achieve highest model performance given the character of the data and unavailability of large training datasets for this task. Moreover, this work exemplifies the application of ML methodology to social media data, which can be particularly relevant for social scientists working in this area and interested in leveraging the possibilities of machine learning in their research field. This methodology and the obtained results provide a foundation for further in-depth analyses of social media texts in the Swedish language following a data-driven approach.

Place, publisher, year, edition, pages
IEEE, 2021. p. 1-8
Keywords [en]
social media, sentiment classification, stance classification, supervised learning, Swedish text data classification
National Category
Natural Language Processing Social Sciences Interdisciplinary
Research subject
Social Sciences; Computer and Information Sciences Computer Science, Computer Science
Identifiers
URN: urn:nbn:se:lnu:diva-108362DOI: 10.1109/SweDS53855.2021.9637718ISI: 000833296400001Scopus ID: 2-s2.0-85123826996ISBN: 9781665418300 (electronic)OAI: oai:DiVA.org:lnu-108362DiVA, id: diva2:1616667
Conference
2021 Swedish Workshop on Data Science (SweDS), Växjö, Sweden, December 2-3, 2021
Projects
DISAAvailable from: 2021-12-03 Created: 2021-12-03 Last updated: 2025-02-01Bibliographically approved
In thesis
1. Frames of threat and solidarity: Dynamics of media discourse on immigration in Sweden
Open this publication in new window or tab >>Frames of threat and solidarity: Dynamics of media discourse on immigration in Sweden
2022 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

This dissertation aims to analyse media discourse about immigration in Sweden in the last decade. To meet this goal, it uses large-scale textual data collected from various media resources, such as mainstream newspapers, social media (Twitter and Facebook) and an online forum. On the one hand, the dissertation explores how the internal architecture of online media contributes to the formulation of public debate about immigration. On the other hand, this work focuses on an external event represented by the refugee crisis and on the ways in which it intervened with the overall discourse dynamics in the Swedish media. Ultimately, this research aims to understand how these internal and external factors affect the framing and construction of the immigration agenda in Sweden. The methodological framework of the dissertation includes a variety of computational text analysis methods, such as sentiment analysis, topic modelling, word embeddings and machine learning, which helps to gain insight into the content and sentiments of the documents published in the media resources. Text analytic methods are further complemented with social network analysis and the study of communication patterns among social media users.

The main results of the analysis indicate that the refugee crisis played an ambivalent role in the overall dynamics of the immigration discourse. While the analysis results suggest several changes in the interpretative repertoires and sentiment of the media content during the crisis,  it is still questionable if they can be characterised as unique or groundbreaking. As for online social media, this work concludes that they have an ambiguous role in the shaping of public debate on immigration. In particular, the discourse on immigration on social media can be characterised as more negative and prone to the influence of such external events as the refugee crisis. At the same time, even minor changes in the platform architecture can indeed influence the ways in which the immigration discourse is formulated on social media. On the other hand, some of the networked properties of social media, such as clustering or homophily, do not necessarily have a negative or polarising effect, contrary to the predictions of network theory.

Place, publisher, year, edition, pages
Växjö: Linnaeus University Press, 2022. p. 58
Series
Linnaeus University Dissertations ; 438
National Category
Sociology (excluding Social Work, Social Psychology and Social Anthropology)
Research subject
Social Sciences, Sociology
Identifiers
urn:nbn:se:lnu:diva-110565 (URN)9789189460676 (ISBN)9789189460683 (ISBN)
Public defence
2022-04-01, 10:00 (English)
Opponent
Supervisors
Available from: 2022-03-16 Created: 2022-02-22 Last updated: 2024-03-13Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Authority records

Yantseva, VictoriaKucher, Kostiantyn

Search in DiVA

By author/editor
Yantseva, VictoriaKucher, Kostiantyn
By organisation
Department of Social StudiesDepartment of computer science and media technology (CM)
Natural Language ProcessingSocial Sciences Interdisciplinary

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 281 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf