Abstract: The knowledge mapping has became one of the most important methods, the reliability testing of this method were still remaining the comparison between results and classical cases. The study gave a comparison between highly cited paper and uncited papers on three network indicators for the aim to test the reliability of co-word analysis. Three indicators were nodes degree, nodes betweenness centrality and isolated nodes base on bibliographic coupling network according to high frequency keywords，which were the citing paper regression of co-word network. The results showed that though the high frequency keywords involved more highly cited papers, there is a little uncertainty in the nodes evaluation according three network indicators. This paper also gave a discussion on the application of thesaurus and data collection of knowledge mapping.

摘要： 知识图谱分析虽然已经成为科学计量学的主要研究方法之一，但对于利用该方法发现学科生长点的可靠性检验大多还停留在经典案例的印证阶段。本文使用基于高频关键词的文献耦合网络回归共词网络指向的施引论文网络，通过比较高被引与零被引论文的节点度数、中介中心性以及孤立节点数等网络指标特征，考察了共词分析结果的可靠性。研究结果显示，高频词关键词更多地指向高被引文献，但使用节点的度数和中介中心性指标去判断节点价值具有不确定性。同时，本文也对标准叙词应用与知识图谱数据采集问题进行了讨论。