摘要
当前e-Learning作为一种重要的教学方式,对其个性化的要求正在日益提高。针对协作过滤推荐是当今应用最为普遍的个性化推荐算法,而数据的稀疏性和算法的可扩展性一直是协作过滤算法所面临的两大问题,提出基于资源类的时间加权协作过滤算法。该算法不仅降低了数据的稀疏性和维度,缩小了目标用户最近邻的查找范围;而且避免了传统推荐算法中推荐值与用户相似度不密切相关的弊端,提高了最近邻的准确度,推荐精度较以往传统算法有明显提高。
Currently e-Learning as an important teaching methods,its individual requirements are increasing. Collaborative filtering is the most common application of personalized recommendation algorithm now, but the sparsity of data and the scalability of algorithm has been a collaborative filtering algorithms faced by the two main issues. Proposed a time weighted collaborative filtering algorithm based on resources category, which not only reduced the data sparsity and dimensionality, and narrowed the area of the nearest neighbor, but also avoided the defects of the traditional recommended method.
出处
《计算机应用研究》
CSCD
北大核心
2009年第6期2107-2109,共3页
Application Research of Computers
基金
国家"十一五"重大科技攻关项目(2006BAH02A24-6)
关键词
个性化推荐
协同过滤
相似度
聚类
personalized recommendation
collaborative filtering
similarity
clustering