Oral history resources are a form of resource that combines personal memory with social memory.The formation of these resources is also a process of social memory reproduction,characterized by distinctive narrative features.Semantic modeling of celebrity historical events within these archives,and the identification of such events at the metadata level,are of significant theoretical and practical importance.This paper optimizes a method for calculating situational similarity of event scenarios in oral history resources based on the construction of a semantic model.Taking the cases of Tsung Dao Lee and the CUSPEA event as examples,it constructs situational units for instance events and extracts situational elements for experiments with the oral history resources in Tsung Dao Lee's special collection.The experimental results show that with a threshold set at 0.6,the accuracy rate is 76.19％,significantly enhancing the efficiency of event identification in such resources.
关键词
口述历史资源/情景相似度/历史事件/自动识别
Key words
oral history resources/situational similarity/historical events/automatic recognition
维普
万方数据