Design And Implementation Of A University Digital Library System Based On Information Recommendation Technology

In big data era,electronic documents shows explosive growth.How to organize the vast amounts of disordered data into knowledge of human society is an urgent problem that digital library become the key to solve.At present the construction of Digital Library in universities is still in the initial stage.The main business of present DL is only about electronic resources management and providing query and load services.However,users often do not know which to choose facing many documents.Information recommendation technology has changed the human way of getting knowledge from traditional active acquisition into passive acceptance,analysis of user characteristics and interest based on the recommendation for interest or information to the user.Under this background,we study the design and implementation of University DL based on information recommendation.Expect for traditional digital library business,we provide information recommendation function to readers.In DL system demand analysis stage,through functional analysis about the existing in digital library system,we model the business of DL based on activity diagram and use case diagram of UML.Finally,we divide Business of University Digital Library into 4 function modules:user information management,electronic document management,query statistics and information recommendation.In order to provide efficient and accurate electronic resource query,we define and model metadata which describes the contents of electronic resources.The metadata is modeled as relational tuple.Two metadata extraction methods are designed separately for different document types.For electronic documents,the extraction method is based on heuristic,which is mainly because of electronic documents has certain regularity can be formulated as a template for automated extraction.On the other hand,for pictures and video,metadata is recorded by artificial way.From the perspective of the needs of digital library scale,using distributed storage method of storage of electronic resources,different file types are stored in different data nodes;for metadata,as well as other clearly defined data,the establishment of E-R model,uses the relational database storage.In the system design and implementation stage,we discuss the system architecture design from 3 aspects:functional organization,architecture and network topology.In J2EE development environment,the university DL system with B/S structure selects Oracle 12C as database and is developed based on JSP technology.In the detail instruction of specific function,we firstly give the overall class diagram,and then based on the flow diagram and sequence diagram the implementation details of the specific function are given.Especially,we provide full-text search function based on Lucene,metadata extraction algorithm based on heuristic rules and information recommendation function based on the collaborative filtering recommendation algorithm.The system test includes two aspects;function test and performance test.Based on black box test method,function test verify the system functions which whether can meet the expected requirement.In the performance testing process,using the simulation software for pressure test,respectively statistics the system response time,upload speed,CPU occupancy rate index.Test results show that,university digital library system based on information recommendation can effectively sorting,storage and retrieval of electronic resources,and provide readers with information query,browsing,downloading and information recommendation service.