中文摘要：

[目的／意义]主题抽取的效果对于信息检索、自动标引、自然语言处理具有重要的价值，提高主题抽取的效果，既能改善检索系统主题检索准确性，又能够帮助学者更加高效地了解文献的主要思想。文章研究并探讨了从引用内容中抽取文献主题的有效性。[方法／过程]选取自然语言处理领域文献为研究对象，利用文献之间的引用与被引用关系抽取引用内容，进行分词并计算权重；将引用内容、全文抽取的候选词进行专家打分评价效果并将其与关键词对比，使用F值评价两种语料库抽取候选词的优劣。[结果／结论]通过专家打分及计算F值，发现引用内容在抽取候选词方面具有明显优势。

英文摘要：

[ Purpose/significance ] The effect of subject extraction has important value on information retrieval, automatic indexing, and natural language processing. Improving subject extraction can not only increase the accuracy of the retrieval system, but also help researchers understand the topics of papers more efficiently. This paper discusses the effectiveness of extracting subject from citations. [ Method/process ] The paper selects papers in natural language processing to do the study. Using the citing and cited relationship among papers to extract citation content, perform word segmentation and calculate the term weight. The paper uses the expert scoring method to evaluate the quality of candidate terms of citation content and the full text and applies the F-measure to evaluate their pros and cons. [ Result/conclusion] By evaluating the candidate terms of citation content and full text by expert scoring method and F-measure, the paper finds that citation contents have obvious advantages in extracting candidate words.