科技论文的研究设计指纹自动识别方法构建与实现
Building and Implement on Automatic Identification Method of Research Design Fingerprint of Scientific Papers
[目的/意义]从科技论文中自动识别与抽取研究设计指纹,能够为科研人员项目设计、研究方法的有效性评估、研究过程问题诊断、研究结果鉴别与评价提供重要的方法论和研究操作支撑。[方法/过程]基于科技论文研究设计指纹的概念模型,提出基于多规则模式混合机器学习方法,设计并实现指纹识别算法,并以数据挖掘领域的期刊文献数据为例,对识别算法的可行性与有效性进行分析验证。[结果/结论]除研究数据与研究趋势外,其他研究设计指纹识别准确率的认可度都基本达到80%以上,覆盖率的认可度,除研究工具与研究数据外,基本达到80%以上。
Purpose/significance] Automatic identification and extraction of research design fingerprint from scientific papers is able to provide researchers with significant methodology and research support for project design, validity evaluation of research methods, problem diagnosis of research process and identification and evaluation of research results.[Method/process]The paper, based on the concept model of research design fingerprint in scientific papers, proposes amulti-rule hybrid machine learning methods to design and implement the fingerprint identification algorithm model and analyze and verify the feasibility and validity of the methodby sample data in the field of datamining.[Result/conclusion] The results show that in addition to the research data and research trends, the recognition accuracy of other research design fingerprint is almost 80%. And the acceptance of coverage, in addition to research tools and research data, is almost 80%.
王茜、钱力、张晓林
自然科学研究方法信息科学、信息技术计算技术、计算机技术
研究设计指纹语义标注知识抽取机器学习
research design fingerprintsemantic annotation knowledge extractionmachine learning
王茜,钱力,张晓林.科技论文的研究设计指纹自动识别方法构建与实现[EB/OL].(2023-08-26)[2025-08-18].https://chinaxiv.org/abs/202308.00415.点此复制
评论