摘要
在目前信息高速发展的时代,个性化推荐作为信息过滤的重要手段,是解决信息超载的最有效方法之一。协同过滤一直是解决个性化推荐比较热门的技术,其主要思想是计算用户之间的相似性或计算项目之间的相似性,然后根据用户或项目之间的相似性对目标用户进行推荐。文章基于协同过滤的思想,提出了一种结合用户评分一致性的单模投影算法,首先在用户与项目的关系二部图中计算用户之间的评分一致性,然后把一致性赋值作为压缩之后的单模投影权值,最后用K近邻找到相似用户并做出个性化推荐。在MovieLens、FilmTrust和Jester等真实数据集上的实验表明,基于评分一致性的推荐算法达到了较好的效果。
In the era of rapid development of information,personalized recommendation,as an important means of information filtering,is one of the most effective ways to solve information overload.Collaborative filtering has always been a popular technology to solve personalized recommendation.Its main idea is to calculate the similarity between users or calculate the similarity between projects,and then recommend to the target users according to the similarity between users or projects.Based on the idea of collaborative filtering,this paper proposes a single-mode projection algorithm combined with user scoring consistency.Firstly,the scoring consistency between users is calculated in the relationship bipartite graph formed by users and projects.Then the consistency assignment is taken as the weight of the compressed single-mode projection.Finally,the method of K-nearest neighbor is used to find similar users and to make personalized recommendations.Experiments on real datasets such as MovieLens,FilmTrust and Jester show that the recommendation algorithm based on score consistency achieves better results.
作者
白源
马浚
刘松华
李泽鹏
BAI Yuan;MA Jun;LIU Song-hua;LI Ze-peng(School of Information Science and Engineering,Lanzhou University,Lanzhou 730000,China;College of Information and Computer,Taiyuan University of Technology,Jinzhong 030600,China)
出处
《广州大学学报(自然科学版)》
CAS
2023年第1期9-16,共8页
Journal of Guangzhou University:Natural Science Edition
基金
国家自然科学基金资助项目(61802158,61762047)
甘肃省自然科学基金资助项目(20JR10RA605)
甘肃省重点研发计划资助项目(20YF3FA024)。
关键词
个性化推荐
协同过滤
单模投影
一致性
K近邻
personalized recommendation
collaborative filtering
one-mode projection
consistency
K-nearest neighbor