分析高影响力论文合作特征对国家、机构、研究人员寻求合适科研合作伙伴以及制定高效合作策略具有参考价值.基于Web of Science数据库,选取论文学科规范化引文影响力值前1％的论文作为高影响力论文,筛得2003-2022年5 997篇中国地球科学领域高影响力论文作为研究对象,基于以合作方式、合作分布以及合作主导性3个维度为表征的嵌入性特征和网络结构特征,从节点个体行为和网络社区结构模型的无监督识别视角识别分析国内外合作网络的结构特征.结果表明:约99％的高影响力论文出自合作,且国际合作比例(53.68％)要高于国内合作比例(46.32％);高影响力论文的国际合作和国内合作均呈现不均衡的集中分布特征,国际合作集中程度更高;中国在国际合作中的主导性不断增强,但与不同国家合作的主导性角色有所不同,国内各机构的合作主导性角色差异也较明显,其中国内外科研合作网络均具有典型的"多核心-边缘"结构,美国和中国科学院分别是国际和国内合作网络中最重要且最靠近几何中心的节点,同时扮演科研合作网络中最重要的中介枢纽角色.为推动中国科研合作高质量发展,应有的放矢地制定个性化的国际合作政策以提高科技合作效率,同时支持鼓励国内合作网络中心圈层的节点机构继续深入开展跨机构科研合作.
Domestic and International Cooperation Characteristics of China's High-Impact Papers:Taking the Field of Earth Science as an Example
As global scientific collaboration becomes increasingly interconnected and interdisciplinary,research collaboration networks have emerged as critical tools for assessing scientific impact and research productivity.International collaboration,often involving multiple countries and institutions,brings diverse perspectives and methodologies to tackle complex problems,thereby driving innovation and enhancing the impact of scientific outcomes.Clarivate's Essential Science Indicators,which include highly cited and hot papers,highlight the novelty and great influence of these studies,serving as key metrics for evaluating research quality and impact.Recently,growing attention has been directed toward the collaborative patterns of high-impact papers,focusing on how various disciplinary,regional,and international collaboration models influence research output through analysis of collaborator relationships.This study leverages the Web of Science database and selects papers in the top 1％of Category Normalized Citation Impact(CNCI)as high-impact papers.A total of 5 997 high-impact papers in the field of Earth Sciences in China from 2003 to 2022 are analyzed.The study examines embedding characteristics and network structural features from three dimensions:collaboration type(domestic vs.international collaboration),collaboration distribution(primacy index and Gini coefficient),and collaboration dominance(leading roles in collaboration).From the perspectives of individual node behavior and unsupervised identification of network community structures,the study explores the structural characteristics of domestic and international collaboration networks.The results indicate that nearly 99％of high-impact papers are products of collaboration,with international collaboration(53.68％)slightly surpassing domestic ones(46.32％).Both domestic and international collaborations display uneven and concentrated distributions,with international collaboration showing a higher degree of concentration.China's leadership in international collaborations has steadily grown,although the nature of its leading roles varies by partner country.Domestically,significant differences exist among institutions in their collaborative prominence.Both domestic and international research collaboration networks exhibit a"multi-core-periphery"structure.In the international network,the United States occupies the most central position,while domestically,the Chinese Academy of Sciences serves as the primary hub,acting as a key intermediary in the research collaboration ecosystem.To achieve high-quality scientific collaboration in China,targeted international collaboration policies should be developed to enhance the efficiency of scientific cooperation.Simultaneously,greater support should be provided to central domestic institutions to deepen cross-institutional research collaborations.
high-impact papersscientific research cooperationcooperation featuresembeddingcooperative network
万方数据
