融合隐语义和邻域算法的兴趣点推荐模型

Synthetic recommendation model for point-of-interest:fusion latent factor and neighborhood-based algorithm

随着众多具有传感功能的智能手机和可穿戴设备的普及,基于位置的服务得到了快速发展,其中基于位置的社交网络(location-based social networks,LBSN)逐渐被大多数人所接受,基于位置社交网络可以为人们提供兴趣点推荐服务。为了提供更加精准的兴趣点推荐服务,提出了一种融合的算法模型。通过隐语义分析算法来充分挖掘用户的历史行为,使用基于邻域的方法结合好友和地理位置等因素,然后在统一的框架中融合这两种推荐方式的结果,实现了对用户行为更好的预测。实验结果表明,提出的兴趣点推荐方法拥有较好的准确率和召回率。

Recently,with the popularity of wearable devices and smart phones which have sensing capabilities,mobile positioning technique is gradually mature. Meanwhile,location-based services have been developed rapidly and the social network included in it,called location-based social networks（ LBSN）,has gradually been accepted by most people. It can provide point of interest recommendation service. In order to provide a more accurate point of interest recommended service,this paper presented a fusion algorithm model. It excavated the previous behaviors of users sufficiently through latent factor algorithm,and used the neighborhood-based algorithm considering other factors such as friends and geographical position. And then fused results of this two recommended ways based on the unified framework which achieved a better prediction of user behavior. The experimental results show that the point of interest recommendation method has better precision and recalling rate.