摘要
在信息检索过程中,因查询词短少而引起的检索歧义性是影响检索效率的主要原因之一,而查询扩展方法和本体扩展方法能有效改善这一问题。提出一种基于本体和局部上下文分析的查询扩展方法:本体扩展根据本体推理规则对短查询词进行推理,得到与查询词有逻辑关联的推理结果集,为查询词加入了标准化的关联信息。局部上下文分析通过对文档库的分析,在与用户查询词最相关的前m篇文档中抽取与用户查询词最相关的n个扩展词,为查询词加入了统计扩展信息。将两部分扩展查询词合并,再通过扩展查询词相关度计算对搜索结果集进行排序。该方法结合了这两种方法的各自优势,从语义角度扩展关键词。实验分析表明,该方法能有效提高检索查全率和查准率。
In the information retrieval process, ambiguity caused by the lack of the retrieval query words is one of the main reasons influencing retrieval efficiency, query expansion and the ontology approach can effectively improve the problem. This paper presents a search extension approach based on ontology and local context analysis： Ontology expansion reasons to the search result set having logi- cal relevance with query words, through ontology inference rules, adding the standardized relevant information into the search words. Local context analysis adds more statistical information for the search words after analyzing the documents, extracting the most relevant n word in the most relevant m files. Join the two part of query expansion word, a set of search results are sorted through the query expansion word relevance computation. The approach in this paper combines the respective strengths of both approaches, extending keywords from the semantic view. Experiments show that this approach can effectively improve the recall and precision.
出处
《控制工程》
CSCD
北大核心
2013年第3期558-561,共4页
Control Engineering of China
基金
国家文化遗产保护科技"十二五"重大项目前期可行性研究课题(20100206)
关键词
本体
局部上下文分析
查询扩展
相关度计算
ontology
local context analysis
search extension
relevance analysis
