Personalized Text Information Of College Institutional Repository Recommended Method Research

With the rapid development of modern network technology,open access has become a popular academic information exchange.The college institutional repository of university scientific research as a kind of academic information exchange platform,it is to get under the background in the open,scientific research personnel at the same time in the open access and use knowledge in deepening,in the face of college institutional repository in a large number of academic information resources,they urgently need a new method to they meet the convenience needs more open access to academic information.How to get the personalized information more quickly and accurately becomes a problem that people can not avoid,which also provides a new application field for the personalized recommendation technology.In this paper,combined with the characteristics of college institutional repository metadata,propose a hybrid recommendation algorithm,this algorithm combines content recommendation technology and collaborative filtering technology based on the application in the college institutional repository of personalized text information recommendation system,convenient for the user to achieve the personalized information recommendation and research users can shorten the time index,can effectively improve the scientific research information the resources of college institutional repository for open access to the huge practical utilization,for the development of university scientific research academic role.The work of this paper includes the following points:First of all,combined with the knowledge base platform of college institutional repository,according to institutional repository metadata features,principle of personalized recommendation technology and related research on information recommendation algorithm,complete all the advantages and disadvantages of common recommendation algorithm in detail,based on content filtering and collaborative filtering recommendation algorithm based on mixed the most suitable for the college institutional repository of text information personalized collaborative filtering recommendation;Secondly,through the analysis of the demand of personalized text information recommendation system,user interest model,clustering algorithm and other personalized information recommendation related to the in-depth study of the main technology,the overall architecture of the hybrid recommendation system was designed in detail,and the structure and process of the system core modules are determined,determined the recommendation strategy,and the design of personalized text information recommendation system of college institutional repository was completed;Finally,according to the research results,a detailed hybrid recommendation algorithm was designed to verify the accuracy of the proposed system.The accuracy and recall rate of the proposed algorithm are verified by the experimental results.Experimental results show that the proposed hybrid recommendation system can match the knowledge base data from the actual institutions of higher education institutions,and the user satisfaction also has a significant improvement,which proves the effectiveness of the proposed system.