Research On Service Model Of University Smart Library Based On The Smart Recommender

The university library has been the service center of teaching implementation and scientific research,the cultural center of academic exchange and cultural communication,the resource center of information transmission and data sharing,undertakes the important social value for readers service.With the idea of ?wisdom of the Earth? proposed,the development of new technology and information explosion appear,the traditional library service has been unable to satisfy the information needs of readers,university library is moving forward to the direction of the smart library.Smart recommender could discover and analyze the behavior of readers actively and initiatively,provide customized information recommender services,to satisfy the increasingly new and diverse demands of users.From the perspective of the smart recommender,this paper try to research university smart library's service model and design the smart service mode to promote the adhesion of the readers of university libraries.In addition,the paper also puts forward some improvement and optimization of recommender algorithm.The main work of this paper is as follows:(1)Investigating 116 of the domestic ?211?universities' smart recommender service by empirical study,site survey and questionnaire.The data suggested that almost all universities have provided some non personalized service content,but the university library's intelligent degree is low,and always depends on the simple recommender function provided by OPAC system.(2)The paper studies university smart library's service model based on smart recommender from four aspects.Including smart retrieval,smart recommender,smart APP and smart micro media service model.(3)Improvement and optimization of recommender algorithm.Considering the readers characteristics of university library,construct the reader feature model,product the generation of candidate nearest neighbor set,resulting in nearest neighbor set,then generating the recommender results,make the recommender results satisfy the readers' preference,in order to improve the accuracy of recommender results.So that the readers have a dependence on recommender service,provide some support for university library's recommender algorithm.