基于概念的文档评价模型

A Concept-based Document Evaluation Model

理解文档的内容和查询的真实意图是提高搜索引擎智能水平的一种有效途径。提出了一种以分析概念及其关系为基础来理解文档、揣摩查询意图的模型。在该模型中,结合用户背景知识构造一个用户概念库,将文档及查询要求转化为概念集,并适当扩充查询要求概念集,最后将两概念集转化为特征向量,计算其相似度,作为文档的评价值。文中详细给出了概念库及必需的概念运算的数学模型。

There exists an obstacle in understanding the retrieval document and in comprehending the request, whose being overcome would effectively improve the search engines performance. This paper presents a concept-based mathematical model to tackle this problem.It includes the concept hierarchy and some concept operations. Based on them, gives a document evaluation schema involving thefollowing stepsas construct a concept base specific to the requester whose background knowledge, his or her interest points, is included, evolve the concept sets from the document and the request and expand the latter one to explicate some hints, translate the sets into two corresponding feature vectors referring the concept relations, evaluat the similarity between them and the degree, the document fulfills the request. In order to implement the schema, it also takes efforts on some basic concept operations.