Research On Accurate Recommendation Service Of Digital Library Based On User Portrait

With the vigorous development of big data information technology,data has become a key output element of all walks of life research and an important source element of digital library innovation and development.The deep development of artificial intelligence technology of big data analysis and machine deep learning leads to significant changes in the process of knowledge discovery and acquisition methods,and the new knowledge technology drives library services to gradually become personalized,refined and intelligent.As a public cultural space,digital library plays an indispensable role in improving readers' knowledge accomplishment and reading level.And library use numerous group,a diversity of content,use of complex,at the same time due to vast amounts of data information choose difficult problem is becoming more and more prominent,the system in order to accurately push information resource for user to read our appetite,let the system independent classification and according to the different characteristics of the user identification information,then according to user needs the personalized recommendation.This dissertation studies the user portrait technology and accurate recommendation related concepts,summarizes the user portrait model construction methods and related technologies.Taking Tianjin University of Technology as an example,a new model of library user portrait is constructed by using questionnaire survey method,K-means clustering algorithm and visual tool word cloud.Combined with collaborative filtering algorithm,user portrait technology is integrated into the field of library precise recommendation service to match the personalized needs of library users.The main research work of this dissertation is as follows:(1)This study defines the concepts of user portrait technology and personalized accurate recommendation,analyzes the basic theories,expounds the modeling process of user portrait technology and the basic elements of personalized accurate recommendation service system,and points out the importance and significance of library for using user portrait technology to realize accurate promotion service.(2)Taking literature statistics and knowledge map analysis as the research method,taking China How Net database as the literature data source,statistically analyze the distribution of document release time and document number,run vosviewer software,and analyze the knowledge map of library user portrait and recommendation system at home and abroad based on the hot research of subject words,so as to reflect the research status and development trend,so as to provide reference for relevant research in the industry.(3)The user data of Tianjin University of technology library are collected.With the help of SPSS software,the K-means clustering algorithm is used to analyze the multi-dimensional data such as users' personal basic information,reading interest preferences and reading evaluation suggestions.The user portrait is constructed by clustering users with similar preferences and attributes,and the word cloud tool is used to visualize the user portrait.(4)Based on the current reading preferences of users,a library accurate recommendation service model is constructed based on the comprehensive use of user portrait and collaborative filtering technology,so as to predict the future reading preferences of users and meet their personalized needs.