摘要
在传统图书推荐过程中,由于用户需求分析的数据集有限,导致用户对推荐图书的满意度较低,为此提出一种以融合多源信息为基础的高校智能图书推荐算法。以多源数据融合为驱动,利用用户规模数据、用户类型数据、用户学龄数据以及用户偏好数据,对用户阅读需求特征分析数据进行扩样处理,并对扩样系数进行限制,从而保留数据的原始特征。根据阅读偏好相似度对扩样后的数据进行分类处理,在同类群体中计算与目标用户历史查阅记录相似度最高的用户,按照借阅时长,将该用户的借阅图书作为最终的推荐结果。在测试结果中,用户对设计算法推荐图书资源的满意度达到了96.0%以上。
In the process of traditional book recommendation, due to the limited data sets of user needs analysis, users’ satisfaction with recommended books is low. Therefore, this paper proposes an intelligent book recommendation algorithm based on fusion of multi-source information. Driven by multi-source data fusion, user size data, user type data, user school age data, and user preference data are used to expand the user reading demand feature analysis data and restrict the expansion coefficient so as to retain the original features of the data.The expanded data are classified and processed according to the similarity of reading preferences, and the user with the highest similarity to the historical reference record of the target user is calculated in the same group. According to the borrowing time, the borrowed books of this user are taken as the final recommendation result. In the test results, users’ satisfaction with the book resources recommended by the design algorithm reaches more than 96.0%.
作者
秦国宾
QIN Guobin(Guangxi Polytechnic of Construction,Nanning Guangxi 530007,China)
出处
《信息与电脑》
2022年第20期94-96,共3页
Information & Computer
基金
2019年度广西高校中青年教师科研基础能力提升项目“‘双一流’建设背景下高校智慧图书馆的研究与实践”(项目编号:2019KY1375)。
关键词
融合多源信息
图书推荐
阅读需求
扩样处理
扩样系数
阅读偏好相似度
fusion multi-source information
book recommendation
reading needs
sample expanding treatment
sample expansion coefficient
reading preference similarity