Stance Detection on Microblog Focusing on Syntactic Tree Representation

摘要: Microblog, especially Twitter, has become an integral part of our daily life, where millions users expressing their opinions towards various target entities. Detecting and analyzing user stances from such massive opinion-oriented twitter posts provide enormous opportunities to journalists, governments, companies, other organizations. However, the short length characteristics frequent use idiosyncratic abbreviations in tweets make this task challenging infer users’ stance automatically a particular target. In paper, we leverage syntactic tree representation detect tweet stance. We devise new parts-of-speech (POS) generalization technique employ hashtag segmentation for effective representation. Then, support vector machine (SVM) classifier with three different kernel functions including subtree (ST) kernel, subset (SST) partial (PT) as base-classifiers. Finally, majority voting count based prediction scheme is employed identify conducted experiments using SemEval-2016 detection dataset. Experimental results demonstrate effectiveness proposed method over baseline state-of-the-art related works.