Boglárka Érsek – The usability of Word Embedding models in social research

2024 Sociology BA Supervisor Renáta Németh, PhD

Boglárka Érsek

In my thesis I explore the usability of word embedding models in social science. My aim is to show what kind of studies researchers have used this method for and how they have used it. In my approach, I focus on the “no code” technique, i.e. I investigate the potential for a researcher who does not know how to program to use the method. In my paper, I will first situate the topic within social science research methods, and then describe the essence of the method and its possible uses. By describing previous research, I will show that the method can be used for both technical and content-related applications, as well as for the critical analysis of algorithms based on linguistic models. In addition, I will demonstrate the applicability to Hungarian texts. In my pilot research, I will demonstrate how the method can be used without programming knowledge by using the online word embedding model WebVectors.

View Thesis