南京理工大学学报（自然科学版）2016，Vol.40Issue(4)：438-443,6.DOI:10.14177/j.cnki.32-1397n.2016.40.04.010
基于图的新闻事件主题句抽取方法
Approach for topical sentence of news events extraction based on graph
摘要
Abstract
News events topical sentence recognition task is a text-based semantic analysis of natural language processing technology. In order to accurately calculate the news text sentences that are most relevant to the topic, this paper proposes a novel approach for topical sentence of news events extraction based on a undirected graph. This paper describes the characteristics of an event trigger Word and sentence extraction named entity and constructs the candidated event extraction templates. This paper,calculates the relationship between candidated event sentences and constructs undirected graphs of event relation ship. Finally,based on the TextRank algorithm,the weight of any vertex in the graph is represented by the weighted sum of the vertex weights, and according to the sorted weights, the event topical sentences are extracted. Experimental results show that the proposed approach is better than TFIDF and event extraction method based on title and that F values are re-spectively 6 . 26% and 2%.关键词
新闻事件/事件主题句/触发词/命名实体/事件关系/无向图/排序/抽取Key words
news events/event topic sentences/trigger word/named entity/event relation/undirected graph/ranking/extraction分类
信息技术与安全科学引用本文复制引用
王雍凯,毛存礼,余正涛,郭剑毅,洪旭东,罗林..基于图的新闻事件主题句抽取方法[J].南京理工大学学报（自然科学版）,2016,40(4):438-443,6.基金项目
国家自然科学基金(61472168) （61472168）
云南省科技厅重点项目(2013FA030) （2013FA030）
云南省教育厅基金重点项目(2015Z022) （2015Z022）
昆明理工大学引进人才科研启动基金项目(KKSY201503007) （KKSY201503007）
