Studying Ideational Change in Russian Politics with Topic Models and Word Embeddings

Abstract

This chapter applies computational methods of textual analysis to a large corpus of media texts to study ideational change. The empirical focus of the chapter is on the ideas of the political role of innovation, technology, and economic development that were introduced into Russian politics during Medvedev’s presidency. The chapter uses topic modeling, shows the limitations of the method, and provides a more nuanced analysis with the help of word embeddings. The latter method is used to analyze semantic change and to capture complex semantic relationships between the studied concepts.

Similarity analysis of federal reserve statements using document embeddings: the Great Recession vs. COVID-19

“…Word embeddings and topic models as computational textual analysis methods are used to describe Russian politics ideational dimensional (Indukaev 2021 ). These techniques develop a research design based on the specificities of the research question and available data.…”

“…These techniques develop a research design based on the specificities of the research question and available data. The study of Indukaev ( 2021 ) describes how the technological innovations and economic development in 2008–2012 changed when Putin replaced Medvedev and selected digitalization as a critical priority by discarding the modernization agenda. This study uses Instagram as the most extensive database in Russian media.…”

Topic modelling discourse dynamics in historical newspapers

“…In this case, however, the authors primarily use LDA as a tool to construct a sub-corpus of relevant articles that was then used for further analysis. Modernization was also an issue in the study of Indukaev [12], who uses LDA and word embeddings to study changing ideas of technology and modernization in Russian newspapers during the Medvedev and Putin presidencies.…”

“…Only Bunout [5] is a clear case of topic instrumentalism. All the other studies depart from some sort of realist position, and attempt to grasp policy shifts, ideas, discourses or framings of topics through topic models, but end up with correctives of some kind by highlighting the interpretative element [21,31], by deploying formal evaluation by historians [9] or by using other quantitative methods to fine tune the results [12]. The interpretative aspect seems especially important when it comes to deciding on what researchers use the topics study as they can reasonable relate to historical discourses, the semantics of related words, or simply ideas.…”

“…This means that any historical analysis using DTM requires a component of historical interpretation of not only topic coherence, but also topic coherence over time. Here, relying on word embeddings like in [12] can help, but this is primarily a task for evaluating the topics.…”