摘要
[目的/意义]以学术论文代表作遴选和评价为目的,探索论文创新性定量测度方法,为代表作评价提供定量数据和客观证据。[方法/过程]对创新概念进行解构,将创新测度指标分为差异度和收益度;使用引用句更客观、丰富地表征文献内容和创新点;将词频统计与词向量相结合,将文本内容与引文网络相结合,提出一种面向代表作评价的学术论文创新性定量测度方法。[结果/结论]对2018年诺贝尔生理学奖获奖论文及其相关论文进行比对实验。结果表明,与被引频次、仅基于文本内容和仅基于引用网络的创新性测度方法相比,该方法更加合理有效,能够更好地适应学术论文代表作的创新性测度要求。
[Purpose/significance]For the purpose of selecting and evaluating representative academic papers,the paper explores an innovation quantitative measurement method so as to provide quantitative data and objective evidence for evaluation.[Method/process]The paper proposes an innovation quantitative measurement method for academic papers,wherein we deconstruct the concept of innovation and build the measurement index both on the degree of difference and profitability;employ the quoted sentences so as to represent the content and innovation of academic papers objectively and abundantly;and take the advantage of multiple methods:the word frequency statistics and the word vector,the text content and the citation network.[Result/conclusion]A comparison experiment conducted on the 2018 Nobel Prize Physiology paper and related papers shows that compared with the method based on citation frequency,text contents and citation networks,the method proposed is more reasonable and more effective in innovation measurement oriented to the evaluation of representative academic papers.
出处
《情报理论与实践》
CSSCI
北大核心
2022年第7期81-88,共8页
Information Studies:Theory & Application
基金
中国科学院文献情报能力建设专项“科研态势感知数据体系和服务平台建设”的成果之一,项目编号:Y8ZG071。
关键词
创新性
差异度
收益度
定量评价
代表作评价
学术论文代表作
degree of innovation
degree of difference
degree of profitability
quantitative evaluation
the evaluation of representative works
representative academic papers
