Research On Key Technology Of Institutional Repository System For Scientific Research Knowledge System

With the development of economic society,the division of labor in each field is becoming more and more detailed,and scholars are increasingly playing a guiding role in all aspects of our lives.Obtaining a variety of expert information,such as research fields and other information,to be processed,and then on this basis,building an Institutional Repository which includes the knowledge of experts is also attracting more and more attention.Institutional Repository can provide great convenience for individual and research institutions who need the experts' knowledge to some extent.The main work of this paper is to use the open source DSpace system to design and implement a unit of Institutional Repository,and to propose a data center-based Institutional Repository data conversion algorithm.This work mainly includes:(1)This paper investigates the management techniques of Institutional Repository.First of all,this paper focuses on the Institutional Repository software which called DSpace System in detail.It analyzes the three-tier architecture of the DSpace system and the functions of each layer,and disusses DSpace System's features in various ways including the data model,metadata,etc.From that it get the advantages and disadvantages of the DSpace System in the practical application.Secondly,this paper introduces the information extraction technology involved in the management of Institutional Repository.It analyzes the classification of information extraction technology,and then obtains the application prospect of information extraction technology in the Institutional Repository.Finally,this paper introduces the algorithms of the same name in detail and mainly analyzes the supervised algorithm of the same name and unsupervised algorithm of the same name.(2)This paper puts forward the design of the Institutional Repository platform model based on DSpace System,and then form the institutional knowledge base system for the scientific knowledge system.In the construction of the Institutional Repository,firstly,this paper presents a retrieval model architecture which includes the part of tagging,the part of speach standardization,stop word processing,index establishment,word meaning disambiguation and extension,word filtering and result retrieval.Secondly,this paper puts forward the data batch import algorithm,which solves the problem of mass data format fusion of large-scale data automation batch import mechanism Institutional repository.Finally,this paper explores,evaluates and displays the expert knowledge in the Institutional Repository.(3)Based on the Institutional Repository system's data,this paper presents a data center transformation algorithm composed of three different algorithms,which are named processing algorithm,classification algorithm and the algorithm of the same name.This algorithm solves the problem that the data of the organization's knowledge base needs to be stored from the document-centric storage management to the data of the storage management of the author and the expert as the center.In this paper,the algorithm of the name disambiguation in the data center conversion algorithm is designed to solve the problem of poor performance of the same name.This paper mainly proposes three algorithms: Firstly,the algorithm of conversion algorithm based on the Relation of Authors(CABRA)based on author relationship is proposed to solve the problem of dislocation of the name disambiguation.Secondly,an improved algorithm for CABRA is proposed,Transformation Based on Graph Structure Algorithm(TBGSA),which solves the problem,that different data size impact on the accuracy to a certain extent of the CABRA algorithm.Finally,based on the TBGSA algorithm,this paper proposes a transformation algorithm based on multi-operator framework(TMOFA),which can achieve the goal of fusion of multi-attribute factors.