一种基于HDP模型的主题文献自动推荐方法及应用研究  

Research on the Application and Automatic Topic Literature Recommendation Method Based on HDP Model

作　　者：颜端武[1] 陶志恒 李兰彬[1] 

机构地区：[1]南京理工大学经济管理学院信息管理系,江苏南京210094

出　　处：《情报理论与实践》2016年第1期128-132,共5页Information Studies:Theory & Application

基　　金：国家自然科学基金项目"新研究领域科学文献传播网络生成及对传播效果影响研究"(项目编号:71373124);总装技术基础研究课题"专题科技信息分析与主题演化平台"(项目编号:G20141QB01)的成果

摘　　要：文章研究了一种基于HDP主题模型的主题文献自动推荐方法,并将其应用于专题情报服务领域知识库构建平台系统的开发。该方法在中文分词和词性标注的基础上,进行文献特征的预筛选获得特征词频分布矩阵,通过HDP主题模型生成各文献与文献集的主题分布矩阵,利用余弦相似度计算与时间阈值加权,生成每篇文献的相关文献集并进行主题文献自动推荐。选取雷达领域语料为样本,在面向专题情报服务的领域知识库构建平台系统中进行了主题文献推荐的模块设计和实际应用。This paper researches on an automatic topic literature recommendation method based on HDP topic model, and ap- plied to develop the domain knowledge base platform of subject information service. The method makes the preliminary screening a- bout the feature of literature on the basis of Chinese word segmentation and part-of-speech tagging to get the word frequency distribu- tion of characteristic matrix in the form of features serial number, and input into the HDP topic model to generate documents and va- rious literature topic distribution matrixes. Finally, the paper uses cosine similarity and combines with time threshold to generate re- lated documents of each literature for the automatic topic literature recommendation. The paper takes the field of radar corpus as samples, in order to show the module design and practical application of topic literature recommendation in the domain knowledge base platform of subject information service.

关 键 词：主题模型 文献推荐 专题情报服务 领域知识库 

分 类 号：G254[文化科学—图书馆学]