Influence Dynamics Among Narratives

Abstract

It is widely understood that diffusion of and simultaneous interactions between narratives – defined here as persistent point-of-view messaging – significantly contributes to the shaping of political discourse and public opinion. In this work, we propose a methodology based on Multi-Variate Hawkes Processes and our newly-introduced Process Influence Measures for quantifying and assessing how such narratives influence (Granger-cause) each other. Such an approach may aid social scientists enhance their understanding of socio-geopolitical phenomena as they manifest themselves and evolve in the realm of social media. In order to show its merits, we apply our methodology on Twitter narratives during the 2019 Venezuelan presidential crisis. Our analysis indicates a nuanced, evolving influence structure between 8 distinct narratives, part of which could be explained by landmark historical events.

Publication
2021 International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation
Akshay Aravamudan
Akshay Aravamudan
Doctoral Student of Computer Engineering
Xi Zhang
Xi Zhang
Senior Doctoral Student of Electrical Engineering

My research interests include point process analysis, modeling and optimization.

Georgios C. Anagnostopoulos
Georgios C. Anagnostopoulos
Associate Professor of Electrical & Computer Engineering

I lead the Machine Learning Research Group at FIT.