Dynamic Modeling And Prediction Of H-index

Academic influence is an important manifestation of the contributions of scholars,institutions or countries to science and technology.How to promote academic influence evaluation to be more objective and fair has been one of the important research directions in the field of library and information.In 2005,Professor JE.Hirsch proposed a new academic impact evaluation index,the h index,which is called a landmark innovation in the field of academic impact evaluation.The h index plays an important role in assessing the academic level of scholars,academic influence of journals,and scientific research influence of institutions.However,for scholars,journals and institutions that are just engaged in academic research,the h index puts them at a disadvantage in comprehensive evaluation.Their short-term academic influence cannot be effectively reflected by the h index.Therefore,studying the time series of the h index can describe the academic influence fluctuations of scholars in different periods,which is more predictive.This article first comprehensively combed the previous research on h-index,and expounded the research status of h-index from four aspects:the improvement and expansion of h-index,the empirical and application of h-index and h-like index,the h-index The relationship between mathematical interpretation and modeling,h-index and the distribution law of information metrology.Then we consider adding the time factor to the h index,and study the time series of h index from the local and global perspectives.Analyzing the change mechanism of the h-index,we can see that the h-index has increased or remained unchanged over time.Based on local analysis and the use of growth curve models,the growth trend of scholars' h-index over the past three years is analyzed to understand the achievements and fluctuations of influence of scholars in differentperiods of their academic careers,and to predict the future h-index of scholars.The results show that the study of the time-series change of h-index is more scientific and suitable for predicting h-index.From the global analysis,the panel data model is used to analyze the time series of the h index from two levels.The research results show that it is more appropriate to predict the h index with a fixed effect model and analyze the time series of the h index.The research significance of this paper lies in solving the problem that the h-index cannot reflect the changes in academic influence of scholars in various periods,and provides more meaningful explorations for the prediction of h-index and the study of dynamic distribution laws.