摘要
当前情报学科的发展目前呈现出多维度、跨学科等特点,而结合个性化信息推荐算法,可为其注入新活力。本文的研究是为了提高个性化信息推荐的效果,解决个性化信息推荐的稀疏性问题,以期可以促进情报学科的新发展,为此,我们引入了社群挖掘概念,得到TO算法,在协同过滤或关联规则推荐之前先对数据进行社团划分,通过对Book-crossing公开数据集的验证分析,并与对照算法相比,我们发现TO算法的准确率和调和度都最佳。
The development of information science is characterized by its multi-dimensional, interdisciplinary nature, and a personalized recommendation algorithm will inject new vitality into it. The research presented in this paper aims to improve the effect of personalized recommendation and to solve the sparseness problem of individual recommendations, to promote new developments in information science. To this end, we import community structure mining into personalized recommendation, which is called the TO algorithm. We mine the community structure of users and items before performing association rule exploration and collaborative filtering. The empirical test based on the Book-Crossing open dataset proves that the precision and F of the proposed algorithm is the best among comparison algorithms.
出处
《情报学报》
CSSCI
CSCD
北大核心
2017年第10期1093-1098,共6页
Journal of the China Society for Scientific and Technical Information
基金
广东省公益研究与能力建设专项"基于社会网络分析的区域协同创新体系研究"(2014B0714021388)
关键词
社群挖掘
个性化推荐
情报学科建设
community structure mining
personalized recommendation
information science construction