Finding Arguing Expressions of Divergent Viewpoints in Online Debates

This work suggests a fine-grained min-ing of contentious documents, specifically online debates, towards a summarization of contention issues. We propose a Joint Topic Viewpoint model (JTV) for the un-supervised identification and the cluster-ing of arguing expressions according to the latent topics they discuss and the im-plicit viewpoints they voice. A set of ex-periments is conducted on online debates documents. Qualitative and quantitative evaluations of the model’s output are per-formed in context of different contention issues. Analysis of experimental results shows the effectiveness of the proposed model to automatically and accurately de-tect recurrent patterns of arguing expres-sions in online debate texts.