Using language models to mine public opinion (10SMDSI_ULM)

Large-scale text data, such as social media or press articles, have become a highly relevant source of information for the analysis of public discourse and opinion. A common interest is to examine how health-relevant topics or, e.g., health policy recommendations resonate with the general public and are discussed in social or conventional media – e.g., whether they are well-received or whether they may instead cause fear or anger. This type of analysis is called 'opinion mining'. An example of this was the preventive measures taken during the pandemic or the vaccination campaigns. In this course, we will focus specifically on classification of large-scale text data using pretrained language models.
We will cover the basics of language models from an applied perspective, explore possible applications for opinion mining, and discuss common challenges (e.g., underlying biases).

MA

1 ECTS

You can find more information about the module here.