用户需求概念图的生成方法研究

摘要

Web information retrieval is widely used today in Internet world, but the accuracy of the retrieval cannot be optimistically satisfied because it breaks up the conceptual relations between submitted keywords. In this paper we start with a kind of user queries in restrained domain, then use methods of rule and statistics comprehensively with search engine as the interface of web corpus source to construct the conceptual graph of query. As the result of query-analysing module, it will guide the following semantic-based searching and raise the relevance of query and returned results. Experiment shows that the method works well and it' s very meaningful in exploring conceptual graph-based semantic retrieval.%网络信息检索在当前互联网社会得到了广泛应用,但是其检索准确性却不容乐观,究其原因是割裂了检索关键词之间的概念联系.从一类限定领域的用户需求入手,以搜索引擎作为网络语料资源的访问接口,综合利用规则与统计的方法,生成查询需求的语义概念图.可将其作为需求分析的结果,导引后续的语义检索过程,提高用户查询与返回结果的相关性.实验结果表明,生成方法是有效可行的,对基于概念图的语义检索有一定的探索意义.