摘要
个性化推荐系统中精确性与多样性似乎是一个鱼和熊掌不可兼得之难题。本文集成基于兴趣社区的用户偏好匹配算法和基于信任邻居的多样性信息推荐算法,构建一个融合精确性和多样性的混合信息推荐模型。实验结果显示:该方法在对推荐结果准确性影响很小的情况下明显提高了推荐列表的多样性。
This paper integrated user preferences matching algorithm which is based on communities of interests and diver- sity information recommendation algorithm based on trust neighbors to design hybrid information recommendation model which merge the attribution of accuracy and diversity. From experiment and evaluation, this model can increase the diversi- ty of recommendations with only a minimal accuracy loss.
出处
《情报科学》
CSSCI
北大核心
2016年第2期65-69,共5页
Information Science
基金
江西省教育厅科技项目(GJJ13290)
关键词
个性化推荐系统
混合推荐
兴趣社区
信任邻居
personalized recommender systems
hybrid recommender
communities of interests
trust neighbors
