摘要
针对目前重启动随机游走推荐算法偏重隐式评分而忽略显式评分的问题,采用监督重启动随机游走算法,使得用户喜爱的项目被访问的概率大于用户不喜爱的项目的概率,从而做出推荐。实验表明,该算法可以有效地提高推荐的准确性。
Aiming at random walk with restart recommendation algorithm mainly for implicit ratings while ignoring explicit ratings, this paper sets random walk under supervision to make recommendation, that makes the probabilities of items which user likes are greater than those of items which user dislikes. Experiment result demonstrates that this algorithm improves the accuracy of recommendation.
出处
《现代图书情报技术》
CSSCI
北大核心
2012年第3期8-14,共7页
New Technology of Library and Information Service
关键词
显式评分
重启动随机游走
个性化推荐
Explicit rating Random walk with restart Personalized recommendation