On using graph centrality measures for DBpedia-based tweet entity linking

Named Entity linking in long text has been well studied in previous works. However, in the context of short text, namely microblogging posts, few works have focused on such form of communication. Micro-posts are short and noisy, thus determining what an individual post is about can be a non trivial task because its highly contextualization and informal nature, which lead to a more complex resolution of synonymy and polysemy problems. In this paper, we present a graph-based system for named entity linking adapted for the context of microblogging posts by assessing the impact of different network centrality measures for the disambiguation task. Our approach relies on the assumption that related entities tend to appear in the same tweet as tweets are topic specific. Also, we address the problem of irregular name mentions. Finally, to show the effectiveness of our system we evaluate it using a real Twitter dataset. Results show that our approach combined with network similarity measures has straightforward results.