摘要
为改进基于位置的社交网络(LBSN)中地点推荐效果,提出一种融合相似性和好友信任的地点推荐算法。将用户划分为目标用户的非好友集和好友集,分别设计基于相似性和好友信任的地点推荐算法。相似性算法基于物质扩散理论,采用好友关系改进;好友信任算法基于社交影响因子。实验结果表明,该算法相比于传统基于用户的协同过滤算法推荐效果明显提高。
To improve the venue recommendation performance in location-based social network（LBSN）,a kind of venue recommendation algorithm combined with similarity and friend-trust was proposed.Users were divided into non-friends set and friends set for targeting the user.The similarity based algorithm and friends＇ trust based algorithm were designed for non-friends set and friends set respectively.The similarity based algorithm was built based on material diffusion theory,and optimized by friendship.The friends＇ trust based algorithm was built based on social influence factor.Experimental results show that the hybrid algorithm achieves better performance than traditional user-based collaborative filtering algorithm.
出处
《计算机工程与设计》
北大核心
2016年第8期2120-2124,2131,共6页
Computer Engineering and Design
基金
国家自然科学基金重点基金项目(71331002)
教育部人文社会科学研究规划基金项目(15YJA630010)
关键词
地点推荐
好友
信任
相似性
协同过滤
venue recommendation
friends
trust
similarity
collaborative filtering
