基于GPS轨迹的用户兴趣点及频繁路径挖掘研究

用户在一定区域内的长期移动行为与某些物理位置相关,如家和办公室等.为了从用户的日常移动轨迹中找出这些位置与轨迹的关系,本文提出一种新的基于移动位置扰动的端点划分与聚类方法.该方法首先将GPS轨迹划分为成组的线段集,然后运用聚类方法将地理上相似的端点进行聚类以检测用户的个性化兴趣点,最后利用原始轨迹来映射个性化兴趣点之间连接关系,生成用户在区域内的移动路径网络.理论与实验结果表明,本文提出的方法可有效地从大量GPS轨迹中发现用户的频繁路径及个性化兴趣点,从而能够为其提供更好的个性化地理信息服务.

Research on GPS-trajectory-based personalization POI and path mining

In this paper, we aim to mine the interesting locations and the frequent travel sequences in a given geo-spatial region, by taking into account the users' historic travel experiences as well as the correlation between locations. First, a new partition method is proposed to divide the trajectories into a set of line segments (in which contains the stationary moving sequence), the start and end points of those line segments are collected as characteristic points. Then clustering method is introduced to cluster the geographical-similar endpoints into groups for detecting fixed territories. Finally an abstract path network is generated to show the linkage relations between the mined fixed territories. The proposed method can be used to detect a user's frequent movement paths as well as fixed territories for better personalized geographical service.