基于改进的用户协同过滤算法的高校个性化图书推荐系统  

College Personalized Book Recommendation System Based on Improved User Collaborative Filtering Algorithm

作　　者：刘佳奇 王全民[1] LIU Jiaqi;WANG Quanmin(Computer Science Academy Beijing Industry University,Beijing 100124)

机构地区：[1]北京工业大学计算机学院,北京100124

出　　处：《计算机与数字工程》2020年第10期2458-2461,2479,共5页Computer & Digital Engineering

摘　　要：基于协同过滤算法是个性化推荐中最基础也是应用广泛的算法,传统的基于用户的协同过滤算法存在用户冷启动,推荐热门占比过高,推荐精度不够高等一系列问题。论文基于现有的协同过滤算法存在的问题,结合高校图书借阅的特点,对这些问题做了改进。通过计算用户相似度来为相似的学生互相推荐对方没有借阅过的图书是基于用户的协同过滤算法的思想,相似度的计算方法和相似用户之间推荐商品的方式是决定推荐效果的核心因素。论文通过学生上网日志聚类和学生的学院信息结合用户的图书借阅记录做用户相似度计算,彻底解决了用户冷启动问题,同时提升了推荐的精度。基于时间衰减的用户兴趣模型可以更加侧重用户近期的借阅兴趣,提升了推荐的精准度。同时也通过清洗推荐样本库的方式解决了个性化推荐中热门占比过高的问题。通过对改进的用户协同过滤算法与传统的用户协同过滤算法进行实验对照,论文的方案在各项指标上都取得了明显的效果。The collaborative filtering algorithm is the most basic and widely used algorithm in personalized recommendation.The traditional user-based collaborative filtering algorithm has a series of problems such as user cold start,high recommended proportion,and high recommendation accuracy.Based on the existing problems of collaborative filtering algorithms,this paper combines the characteristics of college library borrowing to improve these problems.By calculating the user similarity to recommend similar books to each other,the books that the other party has not borrowed are based on the user's collaborative filtering algorithm.The calculation method of similarity and the way of recommending goods between similar users are the core factors that determine the recommendation effect.This paper makes the user similarity calculation through the student online log clustering and the student's college information combined with the user's book borrowing record,which completely solves the user's cold start problem and improves the accuracy of the recommendation.The time-attenuation-based user interest model can focus more on the user's recent borrowing interest,improving the accuracy of the recommendation.At the same time,the problem of too high popularity in personalized recommendation is solved by cleaning the recommended sample library.By comparing the improved user collaborative filtering algorithm with the traditional user collaborative filtering algorithm,the scheme of this paper has achieved obvious results on all indicators.

关 键 词：个性化图书推荐 协同过滤算法 推荐系统 多源数据推荐 

分 类 号：G250.73[文化科学—图书馆学]