摘要
由于社交网络中人物与内容之间错综复杂的关系,如何合理地给用户推荐感兴趣的内容具有十分重要的意义。提出CCVR(Core user for Clustering interesting Vector for Recommend)算法。基于用户的兴趣矩阵,运用改进的K-means算法进行聚类从而推导类兴趣向量,由此预测用户对哪些内容标签感兴趣,从而形成推荐。实验结果证明CCVR算法具有良好的准确性。
How to recommend the interested contents to users reasonably is of great importance because in social networks the relationships between people and content are complicated.We proposed the CCVR （core user for clustering interesting vector for recommend）algorithm. Based on user＇s matrix of interest,we used an improved k-means algorithm for clustering and derived the interests vector of class,thus predic-ted what tags of content the users would be interested,so as to achieve the recommendation.Experimental results proved that the CCVR algo-rithm had good accuracy.
出处
《计算机应用与软件》
CSCD
2015年第10期269-272,共4页
Computer Applications and Software
基金
湖北省教育厅科研基金项目(B20101104)
湖北省重点实验室开放基金项目(znss2013B012)
武汉科技大学科研基金项目(2009xz1)
武汉科技大学大学生科技创新基金研究项目(12ZRC061)
关键词
推荐算法
聚类
兴趣度
Recommendation algorithm
Clustering
Interestingness
