# Contrastive Reasons Detection and Clustering from Online Polarized Debate

## TL;DR

This paper introduces an unsupervised pipeline for detecting and clustering contrastive reasons in polarized debates, improving summarization accuracy by modeling argument facets with a novel interaction-based topic-viewpoint model.

## Contribution

It presents a new unsupervised method combining phrase detection, clustering, and a novel argument facet model for analyzing polarized debates.

## Key findings

- Significant improvement over state-of-the-art in contrastive summarization
- Effective detection and clustering of contrastive reasons
- Enhanced modeling of argument facets in polarized discussions

## Abstract

This work tackles the problem of unsupervised modeling and extraction of the main contrastive sentential reasons conveyed by divergent viewpoints on polarized issues. It proposes a pipeline approach centered around the detection and clustering of phrases, assimilated to argument facets using a novel Phrase Author Interaction Topic-Viewpoint model. The evaluation is based on the informativeness, the relevance and the clustering accuracy of extracted reasons. The pipeline approach shows a significant improvement over state-of-the-art methods in contrastive summarization on online debate datasets.