摘要
【目的/意义】目前针对社会记忆构建的具体方法较少,尚不足以揭示红色记忆中的细粒度关系,利用自然语言处理技术与主题挖掘方法可以挖掘红色档案编研成果中的细粒度关系,有利于红色记忆的重构并对社会记忆构建方法实现有效补充。【方法/过程】本文选取北京香山革命纪念馆红色档案编研成果作为研究对象,对其进行预处理;立足于语义层面及主题层面,通过LDA模型对其展示成果中的五个红色篇章分别深入进行主题挖掘;通过命名实体识别及主题相似度计算的方法抽取其概念、关系与属性,最后构建篇章本体进行可视化展示。【结果/结论】文章利用主题模型对红色档案编研成果进行细粒度挖掘,进而进行本体构建,从而体现其中细粒度关联,实现社会记忆构建,力求实现记忆的映射和更好的呈现以加强档案资源的开发与利用。
[purpose/meaning]At present,there are few specific methods for the construction of social memory,so it is not enough to reveal the fine-grained relationship in red memory.The use of natural language processing technology and topic mining method can excavate the fine-grained relationship in the compilation results of red archives,which is conducive to the reconstruction of red memory and effectively supplements the construction method of social memory.[method/process]In this paper,the research results of red archives compilation of Xiangshan Revolution Memorial Hall in Beijing are selected as the research objects and preprocessed.Based on the semantic level and the theme level,the five red chapters in the display results were mined through the LDA model.The concept,relation and attribute are extracted by the method of named entity recognition and topic similarity calculation,and finally the textual ontology is constructed for visual display.[Results/Conclusion]In this paper,the theme model is used to excavate the fine granularity of the red archive compilation and research results,and then ontology construction is carried out,so as to reflect the fine granularity correlation among them,realize the construction of social memory,and strive to realize the mapping and better presentation of memory to strengthen the development and utilization of archive resources.
作者
陈忻
房小可
孙鸣蕾
CHEN Xin;FANG Xiao-ke;SUN Ming-lei(School of Applied Arts and Sciences,Beijing Union University,Beijing 100191)
出处
《山西档案》
2021年第1期80-87,79,共9页
Shanxi Archives
基金
国家社科基金青年项目“面向社会记忆构建的档案资源检索研究”(18CTQ41)研究成果。
关键词
社会记忆重构
档案编研成果
主题模型
文本挖掘
本体构建
social memory reconstructio
archives compilation and research results
theme model
text mining
ontology construction
