根据用户历史位置记录挖掘用户的日常行为特点
首发时间:2012-01-13
摘要:随着智能手机的普及,GPS已经远远超越了帮助人们用来获取自己位置这一功能,更成为用户提高自己社交效率,分析行为习惯的必要辅助设备。LBS(Location Based Service,基于位置的服务)正是在这一大背景下发展得来的。当前已经有越来越多关于位置信息的算法提出(例如nDCG,MAP算法在LBS中的应用)。本文的一系列分析计算是基于微软研究院等的研究人员所提出的HGSM(Hierarchical-graph-based Similarity Measurement,基于等级结构图的相似度分析)算法。本文算法的目标是通过用户存储的历史位置记录,分析出该用户日常行为习惯,即用户在某个固定的时间段经常去的地点,也就是"时间-地点关系"(Time-Location Relationship,TLR)。得出这一关系之后,可以对用户的出行规律有深入的认识,进而做到商品、好友的相应推荐等。
For information in English, please click here
Mining User Daily Behavior Characteristics Based on User Location History
Abstract:With the popularity of smart phones, GPS not only helps users locate themselves, but also helps users improve their social efficiency and analyze their behavior. The concept of LBS ( Location Based Service) is proposed in this condition. Nowadays, many algorithms (nDCG, MAP) about location history have been proposed. And the algorithm in this paper is proposed on HGSM (Hierarchical-graph-based Similarity Measurement) which is proposed by researchers of Microsoft Research Center. This algorithm, which is known as Time-Location Relationship (TLR) algorithm, is used to mine the relationship between time and location based on users' location histories. Based on the above analysis, friends and commodities could be recommended more effectively.
Keywords: LBS Algorithm Location History User Behavior
基金:
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
根据用户历史位置记录挖掘用户的日常行为特点
评论
全部评论0/1000