摘要
本文对400篇《电子学报》论文中作者标出的关键词及作者提出的创新点词在正文中的句群分布做了统计分析。结果显示关键词0句群分布的概率为31．5％，关键词大于等于0．3的句群分布概率为40．5％，关键词与创新点词用词一致的概率为19．2％；创新点词具有引导特征词和适中的句群分布，由此提出增设创新点词标引的建议。试验表明，该方法不仅专指度高，且有利于创新点的知识单元发现。
The keywords and innovation idea words of 400 papers in （Chinese Journal of Electrnnics）which were put forward by authors in their paper, are statistical analyzed. The analysis result shows that the probability of literatures that the keywords in text have o sentence distribution is 31.5%, and the probability of literatures that the ration of the sentences which include keywords to all sentences in text more than or equal to 0.3 is 40.5%, the probability of literatures that their keywords and innovation idea words are consistent is 19.2%. The innovation words brought forward by authors have obvious lingual structural feature and steady moderate sentences distribution, so a new method of increasing innovation identifier. The experimentations show that the new method can not only decrease error recognizing rate but also open out innovation point, and is a new way to realize knowledge element indexing and knowledge discovery based on innovation point.
出处
《情报学报》
CSSCI
北大核心
2007年第1期50-55,共6页
Journal of the China Society for Scientific and Technical Information
基金
本文为国家自然科学基金项目（70373046）研究成果.
关键词
关键词
统计分析
创新点词标引
句群分布
statistical analysis, innovation idea words indexing, sentence group distributing
