摘要
社会选择和社会影响是在线社交网络社群形成的两个主要因素,如果能有效对网络社群中用户和群体进行分类,就可以采取不同的群推荐策略,实现群体满意最大化。利用偏好对表示群用户偏好,利用矩阵分解和贝叶斯个性化排序方法,考查社会选择和影响对用户偏好的影响程度,实现群用户和群体的分类,进而提出2种群推荐策略。最后通过2个数据集的实验验证,表明本文提出的基于用户和群体分类的群推荐策略是有效的。
Social selection and social influence are the two main factors in the formation of online social network community. If classification of users and groups in network community effectively,it can take different strategies for group recommendation,so can realize the community satisfaction maximization. The paper used pairwise to express user preference preference,used matrix factorization and Bayesian personalized ranking method to judge social selection and influence on user preferences. And it realized the classification of users and groups,Then proposed 2 group recommendation strategies. Finally,the experimental results of 2 datasets showed that the proposed group recommendation strategy based on classification of users and groups is effective.
出处
《现代情报》
CSSCI
2018年第1期92-99,共8页
Journal of Modern Information
基金
国家自然科学基金项目"基于模体挖掘面向在线社交网络中虚拟社区的群推荐系统研究"(项目编号:71371062)
安徽省教育厅人文社会科学重点项目"在线社交网络社区形成机制对企业社会化营销策略的影响研究"(项目编号:SK2015A234)
关键词
社群分类
群推荐
社会选择
社会影响
推荐策略
community classification
group recommendation
social selection
social influence
recommendation strategy