一种基于评分矩阵局部低秩假设融合地理和文本信息的协同排名POI推荐模型  

Synthetic rank-oriented POI recommendation algorithm based on local collaborative ranking exploiting review information and geographical information

作　　者：孙琳[1] 罗保山[1] 高榕[2] Sun Lin;Luo Baoshan;Gao Rong(School of Computer,Wuhan Vocational College of Software&Engineering,Wuhan 430205,China;School of Computer,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉软件工程职业学院计算机学院,武汉430205 [2]武汉大学计算机学院,武汉430072

出　　处：《计算机应用研究》2018年第10期2980-2986,共7页Application Research of Computers

基　　金：武汉市科技局应用基础研究计划资助项目(2015011701011616)

摘　　要：针对目前LBSN中,用户只对少数兴趣点进行签到,使得用户签到历史数据及其上下文信息(如评论文本)极其稀疏,同时传统的评分推荐系统只考虑用户和评分二元信息,具有一定的局限性。为此,提出一种基于评分矩阵局部低秩假设的局部协同排名兴趣点推荐算法。首先,假设用户—兴趣点矩阵在由用户—兴趣点对所定义度量空间中某些邻域内是低秩的;其次,对于地理信息建模采用一种自适应二维核密度方法;然后,对于文本信息利用潜在狄利克雷分配模型挖掘兴趣点相关的文本信息建模用户的兴趣主题;最后,基于局部协同排名模型将兴趣点的地理信息和评论文本信息有效融合。实验结果表明,该模型的性能优于主流先进兴趣点推荐算法。Since the user only check-in a few POIs in LBSN(location-based social networks),so that the historical data of users and its context information(such as review information,geographic information and so on)are extremely sparse,while the traditional recommendation system only consider the user and score binary values,which creates a severe challenge.To cope with this challenge,this paper proposed to utilize local collaborative ranking(LCR)method based on the assumption of locally low-rank rating matrix model for POI recommendation.Firstly,this method assumed that the rating matrix was low-rank within certain neighborhoods of the metric space defined by(user,POI)pairs,which assumed the user-POI matrix was locally low-rank instead of globally low-rank.Secondly,it modeled a personalized check-in probability density with an adaptive bandwidth over the two-dimensional geographic coordinates for each user.Thirdly,it exploited an aggregated latent Dirichlet allocation(LDA)model to learn the interest topics of users and inferred the POIs by mining textual information associated with POI and generated interest relevance score.Further,this paper exploited probabilistic matrix factorization model(PMF)to integrate the review and geographical for POI recommendation to increase the density of local matrix and improve the accuracy of recommendation.Experimental results show that the proposed method outperforms other state-of-the-art POI recommendation algorithms.

关 键 词：局部协同排名 主题相似性 地理偏好 兴趣点推荐 基于位置的社交网络(LBSN) 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]