Research On The Recommendation Of Accurate Reading Based On The Portrait Of University Library Users

University libraries,as the largest physical resource pool of knowledge sources for teachers and students and librarians,play a vital role in improving readers' cultural literacy and reading ability.University library users' behavior is complex,with large crowd,and diverse types.To accurately recommend resources to users,librarians need to understand user characteristics and types,and user portraits are powerful tools for understanding user characteristics and types.By acquiring the user data of university libraries and extracting the user portraits in which the features are clustered,it not only reproduces the picture of the same type of readers,but also is the basis for mining readers' needs and values,performing user segmentation,implementing accurate recommendations and other activities.From the perspective of university libraries,this research uses four methods:literature research,questionnaire survey,statistical analysis and empirical research.By collecting,identifying,and collating existing literature,designing questionnaires,collecting individual users' basic information and reading behavior data of college libraries,this article does theoretical and empirical research on user portraits and give accurate reading recommendations of college library users.The research focuses on the construction of user portraits and accurate reading recommendations of college library mainly from the following aspects:(1)For users,definitions of user portrait and accurate reading are give,the basic theories of portrait and accurate reading recommendation are analyzed,,the components of the user portrait construction process and accurate reading recommendation service system are introduced,and the necessity of accurate reading recommendation using user portrait technology is analyzed.(2)A questionnaire is designed to collect and process user data in college libraries and divide the data into five dimensions:user basic information,user reading environment,user reading behavior,user reading interest and user reading evaluation feedback.(3)The K-means algorithm of WEKA technology tools is used to cluster the data in each dimension to construct a user labeling system and construct user portraits by dividing user groups.(4)The behavior characteristics of users and the existing problems of reading recommendations are analyzed,and seven aspects of the system components of precision reading recommendations are considered,which are:reading recommendation objects,reading recommendation methods,reading recommendation literature resources,reading recommendation evaluation feedback mechanism,reading recommendation infrastructure,Reading recommendation organizations and reading recommendation talents.,By considering these aspects,the strategies of university libraries for reading recommendations and solutions to existing problems are put forward,which provide some meaningful references for the reading recommendations of college libraries.