On October 9th we will have a FullCICS discussion session open to the entire community, organized by PoliCICS. On this occasion, we will have an international guest from Leiden University. Bastián González Bustamante will present his work ‘Unpacking the Unpredictable: From Machine Learning to Generative AI for Detecting Crises in Presidential Democracies’, hosted at OBPEX: https://obpex.com/enlace-udp.
Here are the details of the presentation:
“The study expands on the data collection from my PhD at the University of Oxford (methodological article: https://doi.org/10.1177/08944393251315917), scaling processing from traditional text-as-data approaches to contemporary generative models.
The main objective is to detect and classify critical events, such as social unrest, economic crises, natural disasters, and media scandals, in news archives, link them to cabinet outcomes, and quantify how crises translate into reorganizations in presidential systems. The workflow combines event extraction, multi-coder annotation to build a gold standard, and comparison of LLMs with established classifiers.”
Event Poster
You can view or download the event poster here:
