LBSN中基于活动区域划分的元路径兴趣点推荐

Recommendation of Meta Path Interest Points Based on Active Region Partition in LBSN

基于位置的社交网络(Location Based Social Networks,LBSN)的相关服务推荐越来越多,而兴趣点(Point Of Interest,POI)推荐作为LBSN相关服务中的一项个性化推荐也备受关注,越来越多的学者投入研究。目前,各种基于位置的推荐算法层出不穷,但由于LBSN中的数据极度稀疏的原因,导致许多算法推荐精度不高,本文提出了一种基于用户活动区域划分的元路径推荐算法。首先,根据用户签到以及点评的地点呈现区域性,将用户活动区域分为频繁活动区域和不经常活动区域,根据LBSN结构特征构建用户-活动区域和活动区域-兴趣点之间的二分图模型,其次引入元路径,计算从用户到兴趣点的实例路径的关联度,最后根据关联度大小生成推荐列表。结果表明,该算法较传统的LBSN推荐算法有更好的推荐效果。

Location Based Social network(Location-based Social Networks,LBSN)recommended related services more and more,and the Point Of Interest(Point Of Interest,POI)is recommended as a personalized recommendation in LBSN related services also,much attention has been paid more and more scholars research on.At present,all kinds of location based recommendation algorithm emerge in endlessly,but due to data in LBSN extremely sparse,caused many recommended precision is not high,this paper proposes a meta-path recommendation algorithm based on user activity area partition.First of all,according to the users to sign in and comment on the location of the present regional,frequent user activity area can be divided into active area and not often activity area,according to the characters of LBSN structure building user interest points-activity area and activity area-the dichotomy between graph model,then introduce meta-path,calculated from the user to an instance of an interest point path correlation degree,according to the size of the correlation generated recommended list.The results show that this algorithm has better recommendation effect than traditional LBSN recommendation algorithm.