摘要
基于LBSN(基于位置的社交网络)中数据的地理和社交属性,结合用户轨迹和好友关系,有助于提高不确定轨迹聚类挖掘的效率。根据LBSN用户的好友关系特征,引入评分函数,对用户影响力进行排序,找出其中的活跃用户;在传统路网子轨迹匹配和对签到数据清理的基础上,加入子轨迹匹配准确性监测,并存储活跃用户匹配成功的路段,进而减少路网匹配时间。最后综合R*树的空间索引机制和DBSCAN聚类算法对城市内的热点路径进行挖掘。理论分析和实验表明,相比于已有方法,改进的的聚类挖掘方法在LBSN环境中的时间效率和准确性都有较大的提高,且有较好的可伸缩性。
The data in LBSN （location-based social network） have geographical and social attribute. It is helpful to improve the efficiency of uncertainly trajectory clustering mining combined with user＇ s trajectories and friendship. This paper presented a ranking function based on the feature of friends relationship to sort user＇ s effect and find the active users. Meanwhile, it in- troduced accuracy detection of the road network sub-trajectories to the process of network matching based on data reduction. Moreover, it stored the active users＇ correct matching ways to reduce the time complexity. Finally, it mined hot routes within the city by taking into account both R＊ tree spatial index mechanism and DBSCAN clustering algorithm. Theoretical analysis and experiment results show that compared to the existing method, the method has better stretchability, can get clustering result more accurately and efficiently in the LBSN environment.
出处
《计算机应用研究》
CSCD
北大核心
2013年第8期2410-2414,共5页
Application Research of Computers
基金
重庆市自然科学基金资助项目(CSTC2012jjA40014)
重庆邮电大学博士启动基金资助项目
