基于概念图的信息检索的查询扩展模型

针对传统的基于关键词匹配的信息检索存在的查全率和精确率不高的问题，提出一种基于概念图匹配的查询扩展方法：一方面通过知网对用户查询的词或者句子进行扩展后，将用户查询和文档生成概念图；另一方面利用概念图的不完全匹配和语义相似度的计算方法计算概念图的相似度，以提高检索效果。实验结果表明该方法取得了良好的效果。

Query expansion model based on concept graph information retrieval

One query expansion method, which was based on concept graph, was proposed to solve the low recall and precision rates in the traditional information retrieval methods based on matching keywords. On one hand, words and phrases, which are retrieved by users, can be expanded based on HowNet .Meanwhile, user queries and documents will be transformed into concept graphs. On the other hand, partial matching and semantic similarity based on concept graphs is adopted to acquire similarity between concept graphs, which will optimize the whole retrieval process. This method is proved to be more effective by experiment.