Research On The Classification Of Cross-scientific Journals Based On Community Discovery Algorithms

On November 29,2020,the National Natural Science Foundation of my country officially established the Department of Interdisciplinary Science.Before that,Interdisciplinary became the 14 th discipline category in my country in August this year,and the era of "big science" has come.Based on the whole process of comprehensive scientific research,interdisciplinary science is an inevitable trend of scientific development.With the further in-depth research of any subject,there will always be an insurmountable barrier when it reaches a certain level.There is a trend of interdisciplinary development among basic disciplines to break the shackles of traditional research models.The cohesion between disciplines has made the highly differentiated "specialization" again highly aggregated,and the degree of integration between disciplines has reached an unprecedented level.Few researchers have conducted large-scale analysis of citation relationships between publications,and research is mostly based on the paper level and less on the journal level.Therefore,this paper selects the journal data of MULTIDISCIPLINARY under the ESI classification of the In Cites platform database,and uses the Louvain algorithm to propose a new division method of interdisciplinary scientific journals.Because journal classification is easily affected by objective factors,in order to reduce the deviation of the subject classification system,scientific algorithms are selected for support.In this paper,the Louvain algorithm based on modularity is used for community discovery,which runs fast,and can realize community division of large-scale networks in a relatively short time,and there is no need to specify the number of communities.When the modularity no longer gains,the iteration stops.According to the empirical research on the feasibility of journal classification from the perspective of cross-scientific research,this paper draws the following conclusions:First,Louvain's algorithm has certain practical significance in cross-scientific research.The journal partitioning effect is better,showing a clear knowledge context.If there is a larger amount of data,it will show a more comprehensive academic research map.Second,this paper re-classifies journals,based on the weight table between journals,to a certain extent to reveal the relationship,and the research results show that this research method is feasible.Third,journal evaluation in the perspective of interdisciplinary science needs to be reformed.Facing the needs of academic research and scientific development,it is urgent to promote cross-scientific research.Scientific evaluation is one of the important driving forces to promote the development of interdisciplinary science,in order to achieve the purpose of "promoting construction through evaluation".This paper aims to break through the existing research methods and research models,focus on the knowledge intersection problems that cannot be solved by traditional discipline classification,and clarify the relationship between interdisciplinary scientific journals.Data-driven scientific partitioning is adopted to provide reference for the implementation of the new method of journal classification.