摘要
由于传统推荐算法未考虑项目类型和时间对用户兴趣变化的影响。为此,提出一种基于时间权重和用户兴趣变化的协同过滤算法TACF(time and interests collaborative filtering)。首先,构建用户兴趣分布矩阵,计算用户间的兴趣相似度;其次,引入时间权重函数,计算用户评分相似度;最后,组合两种相似度方法,采用改进的预测评分公式进行计算,提供更加准确的个性化推荐。实验表明,TACF算法不仅能较好地反映用户兴趣的变化,还能提高推荐精确度。
The traditional recommendation algorithm does not consider the impact of project type and time changes on user interest.To solve this problem,this paper proposed a collaborative filtering algorithm called ATCF,which based on user interest change and time weight.First,construct the user interesting distribution matrix and calculate the interest similarity between users.Second,the time weight function introduced to calculate the user score similarity.Final,combined the two similarity methods and the improved predictive scoring formula used for scoring prediction to provide more accurate personalized recommendations.Experiments show that the ATCF algorithm can not only reflect the changes of user interest,but also improve the accuracy of recommendation.
作者
王娜娜
WANG Nana(School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan 232001, China)
出处
《皖西学院学报》
2020年第2期40-45,共6页
Journal of West Anhui University
基金
安徽省教育厅自然科学基金重点项目(KJ2015A111)
2018安徽高校拔尖人才培育项目(gxbjZD15)。
关键词
推荐系统
协同过滤
时间权重
用户兴趣变化
recommendation system
collaborative filtering
user interest change
time weight
