一种基于主题和分众分类的信息检索优化方法  

An Improved Information Retrieval Method Based on Search Topic and Folksonomy

作　　者：李超[1] 

机构地区：[1]解放军南京政治学院上海分院军事信息管理系,上海200433

出　　处：《情报理论与实践》2009年第10期108-110,107,共4页Information Studies:Theory & Application

摘　　要：本文针对目前搜索引擎存在的检索结果缺乏组织导致检准率不高的问题,提出一种基于主题和分众分类的信息检索优化方法。首先对用户检索主题进行获取和表达,然后以社会标签为聚类项,采用向量空间模型实现基于分众分类的文档主题聚类,并将检索结果按相似度和标签"受欢迎度"复合排序,达到提高检索准确率和优化检索的效果。Due to the low pertinency ratio resulting from the lack of organization of retrieval results in current search engines, this paper proposes an improved information retrieval method based on user＇ s search topic and folksonomy. First, the paper obtains and expresses the user＇ s search topic, then, by taking social tags as clustering options, uses SVM model to implement document topic clustering based on folksonomy. Finally, the retrieval results are sorted in a multiple way in accordance with similarity and tag ＂popularity＂ with a view to improving pertinency ratio and optimizing retrieval results.

关 键 词：信息检索 分众分类 向量空间模型 最优化方法 

分 类 号：G354[文化科学—情报学]