View article

Analyzing arguments in user-generated Web discourse has recently gained attention in argumentation mining, an evolving field of NLP. Current approaches, which employ fully-supervised machine learning, are usually domain dependent and suffer from the lack of large and diverse annotated corpora. However, annotating arguments in discourse is costly, errorprone, and highly context-dependent. We asked whether leveraging unlabeled data in a semi-supervised manner can boost the performance of argument component identification and to which extent is the approach independent of domain and register. We propose novel features that exploit clustering of unlabeled data from debate portals based on a word embeddings representation. Using these features, we significantly outperform several baselines in the cross-validation, cross-domain, and cross-register evaluation scenarios.