基于机器学习的高校图书馆用户偏好检索系统研究

Research on User Preference Retrieval System of University Library Based on Machine Learning

[目的/意义]针对大数据环境下高校图书馆检索系统的信息超载问题,提出一种机器学习方法,为用户提供个性化的偏好检索服务。[方法/过程]通过提取用户检索行为大数据中的用户特征,利用监督机器学习方法,在线学习可随用户偏好同步变化的自适应检索模型,预测用户对文献的选择概率,优化检索结果的排序。[结果/结论]设计出用户偏好检索原型系统,介绍用户偏好检索系统工作流程,对比分析系统效果,并对系统进行客观评价。

[ Purpose/significance ] For the information overload problem of traditional retrieval system in university library under big data environment, an online learning method is proposed to provide users personalized preference retriev- al services. [ Method/process] By extracting users＇ characteristics from big data of their retrieval behaviors, and the su- pervising machine learning method, this paper learns an adaptive retrieval model which can synchronize changes with the users＇ preference online, predicts users＇ selection probability for the literature and optimizes the sorting order of the re- trieval results. [ Result/conclusion] This paper designs a user preference retrieval prototype system, introduces the work- flow of user preference retrieval system, makes a comparative analysis on the effectiveness, and objectively evaluates the system.