###
上海电力大学学报:2017,33(6):607-612
←前一篇   |   后一篇→
本文二维码信息
码上扫一扫!
基于上下文关系和TextRank算法的关键词提取方法
(1.上海电力学院;2.浙江华云电力工程设计咨询有限公司)
Keyword Extraction Method Based on Context and TextRank Algorithm
(1.Shanghai University of Electric Power, Shanghai 200090, China;2.Zhejiang Huayun Power Engineering Design & Consulting Co., Ltd., Hangzhou 310000, China)
摘要
图/表
参考文献
本刊相似文献
All Journals 相似文献
All Journals 引证文献
本文已被:浏览 1327次   下载 564
投稿时间:2017-03-09    
中文摘要: 为了从海量文本中高效提取知识,提出了一种基于上下文关系和TextRank算法的中文文本关键词提取方法.首先使用传统方法提取出初始关键词,然后利用互信息筛选出在上下文中对关键词依赖程度大的词,将其作为候选关键词,最后利用TextRank算法计算出最能表达文本主题思想的特征关键词.实验结果表明,与传统方法相比,所提算法在查准率、查全率等相关指标上均有提高.
Abstract:A new keyword extraction method based on context and TextRank algorithm is established to extract the knowledge efficiently from the massive texts.Firstly the algorithm uses mutual information to select the words into candidate key words collection depending on the key words in the context.Then it uses TextRank algorithm to select the words that can express the theme of text.The results show that the algorithm has a higher degree of promotion in precision and recall.
文章编号:20176019     中图分类号:    文献标志码:
基金项目:国家自然科学基金(71203137,61272437);上海市科学技术委员会科研项目(14DZ1207106).
引用文本:
杜海舟,陈政波,钟孔露.基于上下文关系和TextRank算法的关键词提取方法[J].上海电力大学学报,2017,33(6):607-612.
DU Haizhou,CHEN Zhengbo,ZHONG Konglu.Keyword Extraction Method Based on Context and TextRank Algorithm[J].Journal of Shanghai University of Electric Power,2017,33(6):607-612.