Institute for Web Science and Technologies · Universität Koblenz - Landau

Classifying Sentiment Images in Misleading Political News

[go to master theses]

Open Master Thesis - Contact a supervisor for more details!


Images are part of political communication online but are a different vocabulary than language. In this sense, images can function as an extension of news framing and influence reader perceptions. This simple tactic is known from the yellow press and online via “clickbait”. In contrast to sentiment dictionaries, a lexicon of sentimental images is not yet common in political research. This thesis compares language-based results of predicting veracity, popularity, and sentiment to a new set of results produced with the presence of sentimental images. The student must start by surveying the state of image data in political science, social science, and CSS. The idea is to offer these fields a dataset that they want to use and cite. The best way to do so is by identifying a stated need or by expanding on suggestions for future work. This should be the main guideline of the thesis. The student will collect biased news from media bias fact-check[1]. Then, the student will apply a zero-shot learning based classifier to predict emojis from the images [Cappallo, 2015] in news articles and measure the effect of using sentimental images on a media bias dataset that he/she will collect.

  • [Cappallo, 2015] Cappallo, Spencer, Thomas Mensink, and Cees GM Snoek. "Image2emoji: Zero-shot emoji prediction for visual media." Proceedings of the 23rd ACM international conference on Multimedia. ACM, 2015.
  • Rosenthal, Sara, Noura Farra, and Preslav Nakov. "SemEval-2017 task 4: Sentiment analysis in Twitter." Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017). 2017.
  •  Ruder, Sebastian. "An overview of multi-task learning in deep neural networks." arXiv preprint arXiv:1706.05098 (2017).
  • Soroka, Stuart, et al. "The impact of news photos on support for military action." Political Communication 33.4 (2016): 563-582.
  • Hasell, Ariel, and Brian E. Weeks. "Partisan provocation: The role of partisan news use and emotional responses in political information sharing in social media." Human Communication Research 42.4 (2016): 641-661.
  • Soroka, Stuart, and Stephen McAdams. "News, politics, and negativity." Political Communication 32.1 (2015): 1-22.
  • Young, Lori, and Stuart Soroka. "Affective news: The automated coding of sentiment in political texts." Political Communication 29.2 (2012): 205-231.

[1] https://mediabiasfactcheck.com/

Supervisors

  • ibaris@uni-koblenz.de
  • Scientific Employee
  • B 007
  • +49 261 287-2863
  • han@uni-koblenz.de
  • Scientific Employee
  • B 006
  • +49 261 287-2864
  • staab@uni-koblenz.de
  • Professor
  • B 108
  • +49 261 287-2761