摘要
协同过滤由于数据稀疏、冷启动等问题导致了相似用户发现变得困难,从而影响推荐结果的质量。为克服这一不足,提出一种以本体、语义Web为基础的方法,通过构建用户之间的信任网络以及计算用户间的信任权重值,从而增加系统中相似用户发现的数量,改善推荐结果的准确性、可靠性。实例证明,该方法能够帮助推荐系统发现更多的相似用户。
Collaborative filtering is hard to find similar users due to suffering from data sparsity,cold-start problems,resulting in the problem of recommendation accuracy and reliability to some extend.Based on ontology and semantic Web,this paper aimed to alleviate these problems in terms of finding more similar users by automatically building trust network and computing trust weights between users.At last,an example validated the remarkable effects of this approach.
出处
《计算机应用研究》
CSCD
北大核心
2014年第6期1714-1718,共5页
Application Research of Computers
