.png){kind=icon}
返回
Collaborative Filtering with a User-Item Matrix Reduction Technique
.png){kind=icon}
PRE
AI
DOI：10.2753/JEC1086-4415160104.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Collaborative filtering (CF) is regarded as one of the most popular recommendation methods. However, CF has some significant weaknesses, such as problems of sparsity and scalability. Sparsity causes inaccuracy in the formation of neighbors with similar interests, and scalability prevents CF from scaling up with increases in the number of users and/or items. To mitigate these problems, this study proposes a hybrid CF and genetic algorithm (GA) model. GAs are widely believed to be effective on NP-complete global optimization problems, and they can provide good suboptimal solutions in a reasonable amount of time. In this study, the GA searches for relevant users and items from a user-item matrix not only to condense the matrix but also to improve the prediction accuracy. The reduced user-item matrix may reduce the sparsity problem by increasing the likelihood that different customers rate common items. It also shrinks the search space for CF, which ameliorates the scalability problem. Experimental results show that the proposed model improves performance and speed compared to the typical CF model.
Keyword:
Collaborative filtering
genetic algorithms
item selection
recommender system
user selection
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
3.8
论文数:
594
被引数:
2.8K
机构
引用论文
Collaborative filtering using orthogonal nonnegative matrix tri-factorization基于正交非负矩阵三分解的协同过滤算法
PREAI
分享收藏Application of Web usage mining and product taxonomy to collaborative recommendations in e-commerce
PREAI
分享收藏Predicting Missing Ratings in Recommender Systems: Adapted Factorization Approach预测推荐系统中的缺失评级: 自适应分解方法
PREAI
分享收藏Toward global optimization of case-based reasoning systems for financial forecasting.png)
APPLIED INTELLIGENCE
APPLIED INTELLIGENCEIF3.5
PREAI
分享收藏
查看更多内容