Measurement And Hierarchical Cluster Analysis Of Country-Level Citation Strength Based On Web Of Science Academic Big Data

In recent years,with the development of electronic storage technology,as well as the exponential growth of academic data,people have been able to explore the scientific enterprise on a deeper level and higher scale.With the development of storage technology and the improvement of computing power,these data have become much more accessible and analyzable,which has also given birth to a new interdisciplinary field: Science of Science.Science of science aims at mining and analyzing these complex systems composed of academic big data through the theories and technical methods of network science,data science and artificial intelligence,so that people can understand the structures and evolution patterns inside the science from the perspective of data.“Science knows no country,because knowledge belongs to humanity,and is the torch which illuminates the world”,said by Louis Pasteur in 19 th century.With the strengthening of international cooperation and the "open science" movement,science will become more borderless.Though science does not have a "nationality",papers and citation indexes do have.Existing researches demonstrated that a paper tends to cite domestic papers,and papers are disproportionately read in their own country.However,it is not clear that,are countries composing their reference list with national preference,or whether countries' recognition and appreciation for scholar articles have formed community structures,in which the countries belonging to the same community have relatively higher citation strengths,and countries belonging to different communities have relatively lower citation strengths.This paper aims to explore these issues by analyzing the scholar data collected by Web of Science from 1970 to 2016.Due to the different size of scientific output,the exposure rate of citable literature is different in different countries,this paper studies the top 20 countries/regions in terms of the scientific output from the aspects of measuring the citation strength,deteacting the community structures of international citation network,and analyzing the impact of the rapid growth of papers from China on scientific communities: firstly,this paper proposd a null model of random citations based on Z-score,that is,the expected citations received by each country is proportional to their global share of scientific output.Since Z-score is dimensionless,this paper utlizes it to measure the relative standard distance between the actual citation and the expected citation of countries,so that the citation strengths between countries are comparable.After obtaining the matrix of citation strength between countries,considering the dimension and scale of the matrix,this paper adopted the hierarchical clustering algorithm to partition the citation network.Combining the dendrogram obtained by hierarchical clustering and the heat map of citation strength,it is found that there are two obvious groups in the scientific community.Then,this paper proposed six hypotheses for the academic development of China,and discuss the impact of China's rapid development in various disciplines on the scientific community by controlling China's development to the global level.The findings follow:(1)The emergence of "Eastern" and "Western" scientific groups basicly follows people's intuitive cognition of the Eastern and Western world.The only exception is Japan,which is geographically an eastern country but scientificly closer to the West.The "East-West dichotomy" started around 2004 and has been steady for more than a decade.Before that,the international scientific community was an onion-like structure.The Western countries formed the core of the citation network,and the citation strength among each other was very close.Countries in the outer layer,while far from the core,are not strong enough to cite each other to form a new core.This "East-West dichotomy" of societies remains constant when focusing on the citation relationships of papers authored by a single country.(2)The heat maps based on citation strength showed that there is a surplus and deficit in citation strength between "Eastern" and "Western" : "Western" has a surplus in citation strength by "Eastern",while "Eastern" has a deficit in citation strength by "Western".Since the citation is a reflection of the recognition of academic achievements,the citation deficit of "Eastern" by "Western" is also reflected in the international scientific awards.The Nobel Prize,Fields Medal and Turing Award,for example,are rarely awarded to works from "Eastern".Considering China,this paper found that the citation deficit has a significant inhibiting effect on the international citation rate of the menbers in "Eastern".(3)The rapid scientific development in "Eastern" and the relative decline of "Western" may have led to the emergence of this "East-West dichotomy".During last two decades,most eastern countries,such as China,India and South Korea,experienced steady and high level of growth in science.The growing scale of China is particularly significant,so this paper randomly deletes the papers of China to control the developing level of China to the global level.It is found that China's rapid development in specific research fields has largely influenced the emergence of the "East-West dichotomy".To sum up,this paper measureed citation strengths between countries based on a Z-score null model,and utlized hierarchical clustering algorithm to partition the community structure under the null model of citation,thus discovered a potential "East-West dichotomy" in the international scientific community.Finally,through the control experiments,it is found that the rapid development of China in various disciplines plays an important role in the formation of the "East-West dichotomy",which is of instructive significance to the academic development in different disciplines of China.