摘要
协同过滤是大数据推荐系统的重要算法之一,为提高家具推荐效果,文章首先介绍了协同过滤的经典算法,其次对基于图卷积的推荐算法进行梳理并改进,基于收集到的数据集构建了模型,最后选取其他协同过滤经典算法与之进行对比和分析。实验结果表明:相较于基于物品的协同过滤算法模型、基于用户的协同过滤算法模型,基于图卷积的协同过滤算法模型在评价指标上有更好的表现,更适合应用于家具推荐。
Collaborative filtering is one of the most important algorithms of big data recommendation system.In order to improve the effectiveness of furniture recommendations,this paper proposes to improve recommendation algorithm based on graph convolution.Firstly,classic algorithms of collaborative filtering are introduced.Then recommendation algorithms based on graph convolution are sorted out and improved,and a model based on the collected data set is built.Finally,other classic algorithms of collaborative filtering are selected for comparison and analysis.Experimental results show that compared with ItemCF and UserCF models,the collaborative filtering algorithm model based on graph convolution has better performance in evaluation indicators and is more suitable for furniture recommendation.
作者
黄雨洁
李昕
胡基柔
陶卓
HUANG Yujie;LI Xin;HU Jirou;TAO Zhuo(School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China;School of Management,Nanjing University of Posts and Telecommunications,Nanjing 210023,China)
出处
《软件工程》
2023年第11期29-32,共4页
Software Engineering
基金
2022年度江苏省级大学生创新创业训练计划项目(202210293161T)。
关键词
图卷积神经网络
推荐算法
家具推荐
graph convolution neural network
recommendation algorithm
furniture recommendation
