军事事件主题检测与抽取方法
Military Event Theme Detection and Extraction Method
傅浩 1刘姗姗 2李佳蔚 3张志政 3张慧2
作者信息
- 1. 解放军31153部队 南京 210016
- 2. 国防科技大学第六十三研究所 南京 210007
- 3. 东南大学计算机科学与工程学院 南京 211189
- 折叠
摘要
针对主题汇聚功能,为挖掘军事事件潜在语义关联、抽取事件主题知识,提出了一种面向军事事件的主题检测与抽取方法.首先,采用2种聚类策略从不同层面进行事件聚类,以完成事件主题检测;然后,结合词聚类分析与关键词排名算法,给出各合理事件主题表示,并将其作为事件主题知识进一步丰富军事事件知识图谱.
Abstract
Aiming at the topic aggregation function,and to mine the potential semantic association of military events,and to extract event topic knowledge,a theme detection and extraction method for military events is proposed.Firstly,two clustering strategies are used to cluster events from different levels to complete event theme detection.Then,combined with word clustering analysis and keyword ranking algorithm,the reasonable event theme representation is given,and it is used as the event theme knowledge to enrich the military event knowledge graph further.
关键词
知识融合/事件抽取/实体对齐Key words
knowledge fusion/event extraction/entity alignment引用本文复制引用
出版年
2024
NETL
NSTL
万方数据