微型电脑应用2025，Vol.41Issue(3)：210-214,219,6.
基于TF-IDF和面向学科的图书推荐方法研究与实践
Research and Practice Based on TF-IDF and Discipline-oriented Book Recommendation Methods
沈静萍 1张旭 1韩立峰2
作者信息
- 1. 中国石油大学(华东)图书馆,青岛 266580
- 2. 中国石油大学(华东)信息化建设处,山东,青岛 266580
- 折叠
摘要
Abstract
With the continuous development of intelligent library construction,book recommendation becomes one of the impor-tant projects of library intelligent service.Traditional book recommendation methods based on collaborative filtering mainly rely on individual users' reading history and evaluation,without considering the influence of book features on the recommendation results,which has large user-item matrix sparsity and large recommendation deviation.Therefore,this paper analyzes user and book features from the perspective of disciplines,clusters the recommendation objects into different discipline groups,extracts book feature words from book titles and abstracts by training term frequency-inverse document frequency(TF-IDF)algorithm,and constructs book-feature word-feature word weight matrix.This paper obtains the reading preferences of discipline group users from their borrowing history,and recommends books similar to the preferences to achieve precise recommendation for dif-ferent discipline users.The results show that the proposed method has highly accurate and exposure rate for non-popular books,and has good practical significance for deepening discipline construction,constructing academic discipline libraries,and improving the utilization rate of library collection resources.关键词
TF-IDF算法/学科/图书推荐/个性化推荐/阅读偏好Key words
TF-IDF algorithm/discipline/book recommendation/personalized recommendation/reading preference分类
社会科学引用本文复制引用
沈静萍,张旭,韩立峰..基于TF-IDF和面向学科的图书推荐方法研究与实践[J].微型电脑应用,2025,41(3):210-214,219,6.
