事件相似度计算

Event Similarity Calculation

事件间的相似度计算对事件本体的构建起着重要的作用,是基于事件的信息查询、自动问答、自动文摘生成等事件本体应用的前提.本文解决了以下问题,基于给定的事件模型,定义事件间的相似度;根据事件各要素的特征,提出一种结合语法、语义、词语序列、时间关系的事件相似度计算方法.实验数据和模糊分析表明本方法合理、准确.

Event similarity plays an important role in event ontology construction and applications,such as information query,automatic question answering and automatic summarization.In this paper,event similarity is defined based on a given event model,and an event similarity calculation approach is proposed,which combines syntax,semantic,item sequences and time relation,according to the feature of event elements.Our experimental results show that the approach could calculate event similarity comprehensively and accurately.