A Graph-based Approach of Automatic Keyphrase Extraction

• DOI: 10.1016/j.procs.2017.03.087
• Publication Date: 2017-04-08T05:15:08Z
• Authors: Yan Ying; Tan Qingping; Xie Qinzheng; Zeng Ping; Li Panpan
• Keywords: 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

Existing graph-based ranking techniques for keyphrase extraction only consider the connections between words in a document, ignoring impact of sentence. Motivated by fact that word must be important if it appears many sentences, we propose to take full advantage reinforcement andsentences melting three kinds relationships them. Moreover, document is grouped with topics. The extracted keyphrases should synthetic sense they deal all main topics document. Inspired this, topic model into consider. Experimental results show our approach performs betterthan state-of-the-art method on two datasets under evaluation metrics.