摘要
h指数是评价科研人员学术水平、机构科研影响力、期刊学术影响力等的一个综合指标,也是一个非参数统计量。相比之前单一静态的h指数研究而言,研究其随时间变化的动态性更能体现相关对象的学术变化情况,更有预测意义。通过收集整理五个学科970位科研人员的逐年被引数据,并编程计算相应的逐年h指数,研究了动态h指数的一般规律。在分析已有研究的基础上,重点比较并应用逻辑斯蒂曲线模型和龚帕兹曲线模型对科研人员的成长演化过程建模。就所选数据而言,使用逻辑斯蒂曲线模型和龚帕兹曲线模型拟合的效果都很好,R 2值基本达到了0.90以上,且前者的拟合效果好于后者;通过学科差异性比较证实了逻辑斯蒂曲线更能体现h指数的发展演化本质;最后,从增长模型的变化特征细节比较,逻辑斯蒂曲线模型对动态h指数变化规律的拟合更科学。
H index is not only a comprehensive index to evaluate the academic level of scientific researchers,the influence of scientific research institutions and journals,but also a non-parametric statistic.Compared with the previous single static h-index study,the study of its dynamics over time can better reflect the academic changes of the relevant objects,and has more predictive significance.By collecting and sorting out the cited data of 970 scientific researchers in five disciplines,and calculating the corresponding annual h-index by programming,the general law of dynamic h-index was studied.On the basis of the analysis of the existing research,this paper focused on comparing and applying the logistic curve model and gompertz curve model to model the growth and evolution process of scientific researchers.As far as the selected data are concerned,the fitting effect of logistic curve model and gompertz curve model is very good,and R 2 basically reaches above 0.90,and the fitting effect of the former is better than that of the latter;furthermore,the comparison of disciplines proves that logistic curve can better reflect the evolution essence of h index;finally,comparing the changing characteristics of the growth model in detail,the logistic curve model fits the dynamic h-index variation law more scientifically.
作者
牛奉高
梁惠敏
Zhang Jin
Niu Fenggao;Liang Huimin;Zhang Jin(School of Mathematical Sciences,Shanxi University,Taiyuan 030006;School of Information Studies,University of Wisconsin-Milwaukee,Milwaukee,Wisconsin,USA 53211)
出处
《信息资源管理学报》
CSSCI
2020年第1期102-110,共9页
Journal of Information Resources Management
基金
山西省应用基础研究计划资助的山西省优秀青年基金项目“加权共现潜在语义向量空间模型及其在文本主题聚类应用中的惩罚性矩阵分解研究”(201801D211002)
山西省高等学校创新人才支持计划“基于潜在语义的文本信息主题深度聚类研究”(2016052006)
全国统计科学研究项目“高维布尔矩阵降维与分解研究及其在文本聚类中的应用”(2017LY04)的成果之一。