一种改进的协同过滤推荐算法
Collaborative Filtering Algorithm Based on MeanShift Clustering
朱小强 1张琳1
作者信息
- 1. 上海海事大学信息工程学院，上海 201306
- 折叠
摘要
在这个电商及大数据的时代，个性化推荐系统应运而生。由于传统的协同过滤算法存在新使用者、新项目、扩展性以及稀疏性等问题，提出一种基于MeanShift算法的项目或用户聚类，奇异值分解和项目语义相似度的推荐模型，有针对性的对传统的协同过滤推荐系统的不足之处进行改进。实验结果表明，与传统的推荐算法相比，改进算法具有更高的准确性。
Abstract
In this era of electricity and big data, the personalized recommendation is mainly used in e-commerce sites. To solve the traditional collaborative filtering algorithm existing the new user, new project, scalability and sparsity problems, proposes a collaborative filtering algorithm based on MeanShift clustering, Semantic similarity between items and Singular Value Decomposition. To deal with the disadvantages that the traditional collaborative filtering recommendation system facing. The experimental results show that, compared with the traditional recommendation algorithm, the improved algorithm has higher accuracy.
关键词
推荐系统/协同过滤/MeanShift聚类/矩阵稀疏Key words
Recommendation System/Collaborative Filtering Algorithm/Clustering/Singular Value Decomposition引用本文复制引用
出版年
2015
NETL
NSTL
维普
万方数据