Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
SOE: Fachverband Physik sozio-ökonomischer Systeme
SOE 7: Poster
SOE 7.15: Poster
Monday, March 18, 2024, 18:00–20:30, Poster D
The diffusion map technique and its application to political regime types in the 20th century — •Sönke Beier1, Paula Pirker-Diaz1, Matthew Charles Wilson2, and Karoline Wiesner1 — 1Institute of Physics and Astronomy, University of Potsdam, Germany — 2Department of Political Science, University of South Carolina, U.S.
The V-Dem project provides a dataset of dozens of characteristics of political systems, such as press freedom or independence of electoral authorities, for almost all countries and years of the 20th century. But can these high-dimensional data be meaningfully represented in 2- or 3 dimensions and can such a representation help to get insight into the dynamics of political systems?
We have found that the non-linear dimension reduction method "diffusion map" is able to identify a meaningful representation of the data on a 2- dimensional manifold, which is robust over a large parameter range. It sorts countries from autocratic to democratic, also some sub-forms can be recognized. We clarify the influence of the diffusion-map parameters in general, providing insights that are relevant for the application of the technique to other real-word datasets.
In a next step we will use this visualization to analyse different dynamics, such as democratization processes, backsliding, or major disruptive changes.
Keywords: democracy; dimension reduction; diffusion map; V-Dem