摘要
事件语义中的时间语义信息是确定事件边界、支撑事件推理的关键要素,对于业务领域中事件类文本的语义理解和表示有着重要作用。本文根据领域事件中时间语义表示的实际需求提出了一种新的时间语义表示模型GEP(granularityevent-pattern)。该模型在时间粒度、命名时刻与时刻类别等概念基础上,采用代数形式有效地实现了时间语义的表达,利用时刻类别和粒度时刻的动态匹配机制,可灵活、一致地支持多时间粒度下的时间区间表述及周期性事件,表现事件在多时间粒度下的复杂时间语义及其理解过程。通过医疗领域文本理解中的典型案例,验证了模型的可行性和有效性,展现GEP模型在时间语义表达层面的价值,设计了基于GEP模型的GTGD(GEP temporal graph data)图结构数据模型,展现GEP模型在时间语义存储查询方面的应用价值。
The temporal semantics in event semantics is a key element in determining event boundaries and supporting event reasoning,which plays an important role in the semantic understanding and representation of event text.This paper proposes a new temporal semantic representation model named GEP which stands for Granularity-Event-Pattern.Based on the concepts of time granularity,named moments,and time categories,the model effectively uses algebraic forms to to express time semantics.The dynamic matching mechanism of time categories and granular moments can flexibly and consistently support time interval representations and periodic events under multiple time granularity.It can also represent the complex time semantics and understanding process of events under multiple time granularity.Through typical cases in text understanding in the medical field,the feasibility and effectiveness of the model are verified,and the value of the GEP model in temporal semantic expression is demonstrated.A GTGD(GEP temporary graph data)graphstructured data model based on the GEP model is designed to illustrate the application value of the GEP model in temporal semantic storage and query.
作者
李旭晖
冯玉湄
刘柳滟
LI Xuhui;FENG Yumei;LIU Liuyan(School of Information Management,Wuhan University,Wuhan 430072,Hubei,China;Big Data Research Institute,Wuhan University,Wuhan 430072,Hubei,China)
出处
《武汉大学学报（理学版）》
CAS
CSCD
北大核心
2023年第6期809-818,共10页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金重点项目(91646206)
国家社科基金重大项目(21&ZD334)
国家重点研发计划项目(2022YFF0904300)。
关键词
时间语义
语义模式
事件语义
temporal semantics
semantic model
event semantics
