[目的/意义]确定基于引用关系提取关键文献时各种方法的优缺点、适用场合,从而使用户快速捕捉领域重要文献,掌握领域概貌。[方法/过程]基于文献引用关系,从文献被引频次、文献引用网络、文献共被引网络3个角度,结合HistCite、CiteSpace等软件探讨领域关键文献的识别方法,通过同源数据的实际验证,对不同方法进行判别比较。[结果/结论]基于被引频次的方法更适合选择特定领域中哪些文献对总体文献的科学进步产生重大影响角度提取关键文献,对应的关键文献集合呈现非常分散的特性;基于引用网络的方法更适合从特定领域研究动态提取发展过程中的关键文献,对应的关键文献集合呈现非常明显的集中特性;基于共被引网络的方法更适合从特定领域研究基础角度提取关键文献,对应的关键文献集合呈现较强的集中性,且能发现原始采集中未发现的大量关键文献。

[Purpose/significance] Illustrate the advantages and disadvantages between the various methods to extract key documents based on the citation relationships and determine the right situations for different methods allowing users to quickly capture the important documents and grasp the master profile in the field.[Method/process] In this paper, the identification methods of key documents were discussed from three different views such as the times cited, the document citation networks and document co-citation networks based on the citation relations among documents with the help of some programs like HistCite and CiteSpace.The methods were compared based on the same data sets.[Result/conclusion] The method based on the times cited is more appropriate to select the documents in the specific areas which have a significant impact on the scientific progress in the whole set of documents, the distribution of key documents is more divergent.The method based on the citation network is more suitable to select the key documents from the dynamic development in the specific areas, the distribution of key documents is more focused.The method based on the co-citation network is more suitable to select key documents from the research base in the specific areas, the distribution of the key documents is focused and a number of key documents were found out of the range of the original acquisition.