摘要
学术成果的跨界影响力评价在基金立项、科研管理、学科建设和学术评审等领域有着广泛的应用前景。不同于学术深度评价,它是对学术广度交叉规律的认识。其中跨界关系强度矩阵是跨界影响力研究的关键。以人文社会科学学术成果的跨界影响力数据模型为切入点,基于读者、作者通过文献搜索所建立的学术跨界联系,本文探讨了几种跨界关系矩阵模式发现算法,比较了三种前沿统计网络估计方法在跨界影响力上的应用。实证研究表明,这些算法具有稳定和快速的特点,在人文社会科学跨机构和跨学科关系研究方面得到了有益的结果。
Crossover Impact Evaluation (CIE) within academic achievements is required inputs for many scientific evaluation tasks including academic management, funding planning and disciplines development. Different from depth evaluation within discipline, CIE is aimed to reveal the crossover impact between disciplines. The critical problem is the estimation of crossover impact matrices. It is difficult to measure crossover impact directly. In this paper, we explore the relative algorithms within hu- manities and social sciences. There has been recent interest in inferring correlation matrices from link measurements and other more easily measured data. This inference problem can be solved by social network algorithm. This paper explores three approaches to crossover impact matrix estimation. The empirical study demonstrates the robust and fast property and reveals valuable information on interdisciplinary and cross institutes.
出处
《中国人民大学学报》
CSSCI
北大核心
2012年第4期134-143,共10页
Journal of Renmin University of China
基金
中国人民大学科学研究基金(中央高校基本科研业务费专项资金资助)研究品牌项目"人文社会科学成果评价指标体系"(10XNI014)
关键词
学术成果评价
跨界影响
社群发现
关系估计
academic achievement evaluation Crossover Impact Evaluation community discovery
relationship matrix estimation