论文部分内容阅读
随着移动设备和定位技术的广泛应用,基于位置的社交网络(LBSN)通过签到能够记录用户的历史访问轨迹.对LBSN移动轨迹的研究存在用户位置信息复杂且无效数据多、用户社交关系没有充分利用和位置预估计不精确的问题.首先针对虚假签到和无效用户信息提出一种基于泊点频率的去噪算法,其次在基于密度聚类算法的基础上,结合地理位置信息和用户关系,综合考虑用户兴趣相似度及信任度,提出针对位置预估计的组合服务算法.实验结果表明,本文提出的方法能有效改善位置预估计的精确度.
With the widespread application of mobile devices and location technologies, location-based social networks (LBSNs) can track the history of users’ access by checking in. The research of LBSN’s mobile trajectory has complicated and invalid user location information and insufficient social relationships Utilization and location estimation inaccurate.In this paper, we first propose a de-noising algorithm based on the pooled frequencies for the spurious sign-on and invalid user information. Secondly, based on the density clustering algorithm, combined with the geographic location information and user relationship, Considering the similarity of users’ interest and the trust degree, a combination service algorithm for location estimation is proposed.The experimental results show that the proposed method can effectively improve the accuracy of location estimation.