Word-sentence co-ranking for automatic extractive text summarization

详细信息    查看全文

• 作者：Changjian Fang^a ; ^{changj.fang@njue.edu.cn} ; Dejun Mu^a ; ^{mudejun@nwpu.edu.cn} ; Zhenghong Deng^a ; ^{dthree@nwpu.edu.cn} ; Zhiang Wu ; ^b ; ^{zawuster@gmail.com}
• 关键词：Automatic summarization ; Extractive summary ; Sentence scoring ; Pagerank ; Unsupervised learning
• 刊名：Expert Systems with Applications
• 出版年：2017
• 出版时间：15 April 2017
• 年：2017
• 卷：72
• 期：Complete
• 页码：189-195
• 全文大小：585 K
• 卷排序：72

文摘

A principled word-sentence co-ranking model called CoRank is proposed. The convergence of CoRank with matrix notation is proved. A redundancy elimination technique is presented to further improve the performance of CoRank.