基于SVD的档案学主题挖掘
Text topic mining of archives research based on SVD

摘要： 收集2010—2014年国家社科基金档案学领域立项课题,基于课题名称进行分词等预处理,得到词项-文档矩阵,依据词项重要性设计局部及全局权重,组合局部与全局权重,得到词项-文档矩阵权重值。利用奇异值分解SVD进行特征降维,研究在不同维度下近5 a国家社科档案学立项课题研究主题。经过可视化分析得到社科档案学七大研究主题为:非物质文化遗产保护、电子文件管理、数字资源建设及体系、档案信息资源价值与挖掘、档案保护机制、档案馆研究、档案信息安全。
Abstract: The data of National Social Science Fund Project on Archives Field from 2010 to 2014 were collected, the words of the project title are separated, and the term-document matrix was obtained. According to the importance level of the terms, local and whole weight was designed, local weight was integrated with whole weight, which obtained the weight value of the term-document matrix. Feature dimension reduction was implemented by SVD, the recent National Social Science Archives Project themes in different dimensions were studied. Eventually, seven research topics of social science archives were obtained by visually analyzing, which were the intangible cultural heritage protection, electronic document management, digital resource construction, value and research of the archival information resource, archival information protecting system, research of the archives, security of the archival information