Quantitative Research On Long-term Derivation Model In Academic Research Domain

In the academic research,the research work of scientists often publishes articles to show their research progress and results,and these articles will be studied by other scholars in these related domains.Aggregating related articles will form one domain after another,accompanied by the termination of the old domain and the emergence of new domain,thus forming a process of domain derivation and replacement,which may also follow some universal law.Therefore,this thesis is to study this law,that is,based on the citation network of academic research domain,we study the long-term evolution model of the research domain.In order to achieve the above objectives,this thesis constructs the domain-derived space to present the derivation of different domains,combining with domain production dynamics,the dynamic characteristics of domain and ancestral domain are proposed and quantified to analyze the derivative dependencies between domain and its ancestral domain.First,construct a domains-derived dependency statistically validated network,and stage the domain based on the growth rate and start year.Then,combining the structural characteristics of the network,quantifying the four characteristics to study the derivation of the domain.Based on the dynamic characteristics of domain and domain-derived space,the instantiation of long-term domain derived patterns is realized on two datasets.According to the above research process,the conclusions of the new domains are derived from the interdisciplinary effect to the endogenous drive,and most of the new domains are generated as the early influence of the existing disciplines.In summary,in order to achieve quantitative research on the derivation model of long-term follow-up in the domain,this thesis designs and constructs the Domainderived Space,and combining the network structure characteristics,this thesis proposes a long-term derivation model for exploring the domain from multiple directions.This thesis provides a set of methods for data mining and analysis of massive relational data,it is of great significance for researchers to study different relational data and predict the development of the domains.