LBSN中考虑用户交友偏好的好友推荐方法研究
发布时间：2019-04-11 20:45
【摘要】：基于位置的社交网络(location-based social networks,LBSN)大为流行之余,也带来了信息过载问题.好友推荐是所有社交网络必须面临的问题,为了改进LBSN中好友推荐的效果,构建了考虑用户交友偏好的好友推荐模型(friends recommendation considering users'preference,UPFR).从兴趣相似性、距离和熟识度三个属性刻画LBSN中的用户,兴趣相似性属性基于信息熵理论计算、距离属性通过朴素贝叶斯推导、熟识度属性建立在共同好友的基础上.在对三个属性进行集成时,考虑了用户的交友偏好,通过目标用户的好友列表确定各属性的权重,建立了自适应用户交友偏好的好友推荐算法.通过Foursquare上的数据实验证明该算法能取得较优的综合推荐效果.
[Abstract]:Location-based Social Network (location-based social networks,LBSN) is not only popular, but also brings the problem of information overload. Friend recommendation is a problem that all social networks have to face. In order to improve the effect of friend recommendation in LBSN, a friend recommendation model (friends recommendation considering users'preference,UPFR) considering the preference of users is constructed. This paper describes users in LBSN from three attributes: interest similarity, distance and familiarity. Interest similarity attributes are calculated based on information entropy theory, distance attributes are deduced by naive Bayes, and familiarity attributes are based on mutual friends. In the integration of the three attributes, considering the user's preference for making friends, the weight of each attribute is determined by the friend list of the target user, and a friend recommendation algorithm based on the adaptive user's preference for making friends is established. The experimental results on Foursquare show that the proposed algorithm can achieve better comprehensive recommendation results.
【作者单位】： 合肥工业大学管理学院;
【基金】：国家自然科学基金重点项目(71331002) 教育部人文社会科学研究规划基金项目(15YJA630010) 国家自然科学基金面上项目(71571059)~~
【分类号】：TP391.3
本文编号：2456729
[Abstract]:Location-based Social Network (location-based social networks,LBSN) is not only popular, but also brings the problem of information overload. Friend recommendation is a problem that all social networks have to face. In order to improve the effect of friend recommendation in LBSN, a friend recommendation model (friends recommendation considering users'preference,UPFR) considering the preference of users is constructed. This paper describes users in LBSN from three attributes: interest similarity, distance and familiarity. Interest similarity attributes are calculated based on information entropy theory, distance attributes are deduced by naive Bayes, and familiarity attributes are based on mutual friends. In the integration of the three attributes, considering the user's preference for making friends, the weight of each attribute is determined by the friend list of the target user, and a friend recommendation algorithm based on the adaptive user's preference for making friends is established. The experimental results on Foursquare show that the proposed algorithm can achieve better comprehensive recommendation results.
【作者单位】： 合肥工业大学管理学院;
【基金】：国家自然科学基金重点项目(71331002) 教育部人文社会科学研究规划基金项目(15YJA630010) 国家自然科学基金面上项目(71571059)~~
【分类号】：TP391.3
【相似文献】
相关期刊论文 前1条
1 胥皇;於志文;封云;周兴社;;基于LBSN的个性化旅游包推荐系统[J];计算机与现代化;2014年01期
，本文编号：2456729
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2456729.html
