STEntConv: Predicting Disagreement with Stance Detection and a Signed Graph Convolutional Network
This addresses the challenge of understanding polarised online discussions for social media analysis, though it is incremental as it builds on existing graph-based methods.
The authors tackled the problem of predicting agreement or disagreement between social media posts by leveraging user stances on named entities, and their STEntConv model improved disagreement detection performance on Reddit data without needing platform-specific features.
The rise of social media platforms has led to an increase in polarised online discussions, especially on political and socio-cultural topics such as elections and climate change. We propose a simple and novel unsupervised method to predict whether the authors of two posts agree or disagree, leveraging user stances about named entities obtained from their posts. We present STEntConv, a model which builds a graph of users and named entities weighted by stance and trains a Signed Graph Convolutional Network (SGCN) to detect disagreement between comment and reply posts. We run experiments and ablation studies and show that including this information improves disagreement detection performance on a dataset of Reddit posts for a range of controversial subreddit topics, without the need for platform-specific features or user history.