Research And Implementation Of The Key Technologies Of Heterogeneous Knowledge Warehouse Data Integration

University institutional repository is a mainstream application of knowledge warehouse.Since development,it has played an important role in academic communication,electronic publishing,long-term preservation,knowledge management,education promotion,scientific research assessment,sharing data and so on.However,university institutional repositories are independent,which causes that academic resources cannot be shared and difficult to use and communicate.With the demand of network digital literature information resources growing,the independent resource of institutional repository can't meet the demand,the Co-construction and sharing of resources and provision "one-stop" services for users become inevitable.Thus,this project implement the interconnection between institutional repositories to establish centralized data warehouse and provide users with a unified access services for resources.Data replication is adopted to establish the centralized data warehouse,which harvests data in each institutional repository to data integration center.Data harvesting includes metadata harvesting and object data harvest,which uses OAI-PMH protocol and METS protocol to package object data into METS document and embedded into OAI record,thus implement the joint harvest of both of the data.Meanwhile,to improve its flexibility,the project provides two kinds of metadata:etd and etds,both of which define the metadata format of OAI record.In the data harvesting system,each institution acts as data provider,which uses OAICat to provide OAI record;and data integration center acts as data harvester,which uses OAI interface of institutional repositories to harvest OAI record,resolve metadata and object data of resource entry,and save them into data integration center.After harvesting metadata and object data,the project conduct semantic analysis of academic resources of one subject in the data warehouse and get domain keyword library of this subject,based on which implement semantic association between the academic resources.When browsing the academic resources,users can quickly discover new resources by clicking on the associated keywords,thus provides personalized knowledge service for users.