Dynamic Evolution Research On Scientific Research Cooperation Network Based On Academic Preference

Under the background of great science,scientific research cooperation has become the main form of scientific and technological innovation activities.In recent years,the phenomenon of scientific research cooperation has aroused wide concern.The scientific research cooperation network as a complex network theory in the study of cooperative research in the specific application,by mapping the academic preferences of the network evolution rules,from the perspective of dynamic network research cooperation network dynamic evolution process,can effectively grasp the scientific research cooperation network evolution Rule,optimize the structure of scientific research cooperation team.Research papers as the main form of scientific research achievements,extracting the internal characteristics and external characteristics of scientific research papers have become the research methods recognized and widely used by domestic and foreign academic circles.For this reason,this paper we introduce the theory of preference and the theory of network dynamic evolution into the research of scientific research cooperation to carry out the dynamic evolution research on scientific research cooperation network based on academic preference,this paper mainly studies the following several aspects:(1)the evolution of author cooperation weighted network.In this paper the author of scientific research cooperation as the basic elements to build a network of author cooperation,the nodes in the network represent the authors,the size of the nodes represents the weighted frequency of the authors,the edges represent author cooperation relationship,the thickness of edge represents the strength of relationship between authors.Firstly,introduces the basic theories of network evolution and weighted method.On the basis of improving the growth model the dynamic evolution model of author cooperative weighting network based on academic preference is constructed,the author cooperation network is driven by the author's academic preference;Then the network evolution is simulated by setting different parameters and the corresponding network degree distribution is calculated;Finally,the distribution of the network degree in the evolution process of author cooperation weighted network is verified by the cooperation of the sample data,confirm that the dynamic evolution model of author cooperative weighting network is universal and reasonable,and the network characteristics of the core author network in the sample data are analyzed.(2)The Evolution of Academic Community Network.The basic elements of academic community network contains authors,institutions and key words,and the academic community network is divided into homogeneous and heterogeneous network according to the node attributes.The homogeneous network is non-directional network,the heterogeneous network is a directional network.Firstly,the basic theory of the academic community network evolution and the meta-network representation of the academic community are introduced.Then,the evolution model of the academic community based on academic preference is constructed on the basis of the basic growth model.The evolution process of academic community network is based on the author's layer network evolution,the other layers of the network according to the co-occurrence probability rule synchronization evolution;finally,the academic community network constructed by the sample data was visualized.The model can well explain the relationship between the author,the organization and the key words,and reveals the rules of the evolution of the academic community network.(3)Recommendation of scientific research cooperation partner.For researchers recommend cooperation partner with common academic preference can optimize the structure of scientific research team,promote knowledge innovation and knowledge integration.Firstly,this paper introduces three types of co-occurrence relationship based on keyword co-occurrence,co-citation and co-citation,proposes a method to construct the co-authors network based on the weight of the contribution of scientific research,And analyzes the recommendation thought of scientific research cooperation partner based on academic preference;then introduces the vector space representation method of academic preference for researchers,and the modified cosine similarity method is used to transform it into academic preference similarity;finally,we use the sample data to carry on the empirical analysis and use the standard deviation measure the consistency of the results,and the comprehensiveness and accuracy of the results of are measured by recall and precision.The method of weighting the subject of scientific research by using scientific research contribution,can more accurately measure the similarity of academic preferences among researchers,and then be able to recommend more suitable partners for scientific researchers.