基于用户偏好行为数据的数字图书馆资源推荐方法研究  

Research on Resource Recommendation Method of Digital Library Based on User Preference Behavior Data

作　　者：王秀慧[1] 赵治军 WANG Xiuhui;ZHAO Zhijun(Institute of Education Science and Technology,Shanxi Datong University,Datong Shanxi 037009;Information Center of Datong People's Government,Datong Shanxi 037009)

机构地区：[1]山西大同大学教育科学与技术学院,山西大同037009 [2]大同市人民政府信息化中心,山西大同037009

出　　处：《软件》2022年第5期8-12,45,共6页Software

基　　金：山西省哲学社会科学基金项目“基于知识图谱的数字图书馆精准推荐模型构建研究”(2020YJ148)的研究成果之一。

摘　　要：为给读者提供更加精准的推荐服务,提出了一种基于用户偏好行为数据的数字图书馆资源推荐方法。从用户对数字图书馆资源的偏好行为中提取浏览、评论、下载三种数据,结合用户资源评分,构建用户偏好行为数据推荐模型。在确定模型中参数取值以确保模型方法最优推荐精度后,与传统推荐方法进行实验比较。通过实验验证该方法模型在准确率、召回率方面均优于传统的协同过滤方法,有效解决了资源评分数据稀疏性的问题。In order to provide readers with more accurate data recommendation services,a resource recommendation method of digital library based on user preference behavior data is proposed.Three kinds of data are extracted from users’preference behavior for digital library resources:Browsing,commenting and downloading.Combining the user preference behavior data score with the user resource score,the user preference behavior data recommendation model is constructed.After determining the parameter values in the model to ensure the optimal recommendation accuracy of the model method,the experiment is compared with the collaborative filtering method.The experimental results show that the proposed model is superior to the traditional collaborative filtering method in terms of accuracy and recall rate,and effectively solves the problem of sparsity of resource rating data.

关 键 词：数字图书馆 推荐 协同过滤 用户偏好 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论] G250.7[自动化与计算机技术—计算机科学与技术]