Bounded-confidence Cascade Parameter Fitting

Bounded-confidence cascades simulate the spread of ideas on a social network. Fitting these models with social media datasets would let us study the mechanics of online political polarization. However, bounded-confidence model fitting is largely unexplored by the field due to incompatibility between messy, real datasets and the model’s abstract foundations. With the larger goal of model fitting, this project involves developing tools to extract and analyze follower and retweet network structures from Twitter and statistical estimation of the models’ underlying network parameters. A key challenge is adapting complex textual data to the model’s assumption that opinions are single numerical values. To address this, the project employs probabilistic text classifiers to assign opinion scores directly to text data, mainly through fine-tuning logistic regression models and LSTM neural networks on extracted social media datasets.

[Poster]: Zinn-Brook H, and Pang-Naylor K. 2022. Twitter Generated Information Cascades. Poster session presented at: HMC Summer Research & Scholarship Poster Celebration; 2022 September 23; Claremont, CA.