徐乾,,
黄瑞阳,
程晓涛,
吴铮
国家数字交换系统工程技术研究中心 ??郑州 ??450002
基金项目:国家自然科学基金(61521003)
详细信息
作者简介:陈鸿昶：男，1964年生，教授，研究方向为通信与信息系统
徐乾：男，1993年生，硕士生，研究方向为社交网络挖掘、机器学习
黄瑞阳：男，1986年生，副研究员，研究方向为网络大数据处理与分析
程晓涛：男，1986年生，博士生，研究方向为社交网络挖掘、机器学习
吴铮：男，1992年生，博士生，研究方向为复杂网络、网络大数据处理与分析
通讯作者:徐乾 549529376@qq.com
中图分类号:TP393; TP391.4计量
文章访问数:1904
HTML全文浏览量:652
PDF下载量:83
被引次数:0
出版历程
收稿日期:2018-01-30
修回日期:2018-06-11
网络出版日期:2018-06-30
刊出日期:2018-11-01
User Identification Across Social Networks Based on User Trajectory
Hongchang CHEN,Qian XU,,
Ruiyang HUANG,
Xiaotao CHENG,
Zheng WU
National Digital Switching Engineering & Technological Research Center, Zhengzhou 450002, China
Funds:The National Natural Science Foundation of China (61521003)
摘要
摘要:针对现有的基于用户轨迹的跨社交网络用户身份识别算法未考虑用户轨迹中的位置访问顺序特征的缺点，该文提出一种基于Paragraph2vec的跨社交网络用户轨迹匹配算法(CDTraj2vec)。首先将用户轨迹转化为易于处理的网格化表示，并按照一定的时间粒度、距离尺度对原始的用户轨迹进行划分，使用户轨迹中的位置访问顺序特征易于抽取；然后利用Paragraph2vec算法中PV-DM模型抽取轨迹序列中位置访问顺序特征，得到用户轨迹的向量表示。最后通过用户轨迹向量判定轨迹是否匹配。在社交网络BrightKite上的实验结果表明，与基于位置访问频率或者基于轨迹间距离的方法相比，F值提高了2%～4%个百分点，所提算法能够有效地抽取出用户轨迹中的位置访问顺序特征，更加准确地实现了基于用户轨迹的跨社交网络用户身份识别。
关键词:社交网络/
用户身份识别/
轨迹相似度/
Paragraph2vec
Abstract:The performance of trajectory based user identification is poor since the existing methods ignore the order feature of location sequence. To solve this problem, a Cross Domain Trajectory matching algorithm based on Paragraph2vec (CDTraj2vec) is proposed. Firstly, the user trajectory is transformed to the grid representation which is easy to handle. The PV-DM model in the Paragraph2vec algorithm is utilized for extracting order feature of location sequence in trajectory. Then the original user trajectories are divided by a certain time size and distance scale to construct a training sample suitable for training PV-DM model. The PV-DM model is trained by different types of training samples, and the vector representation of the user trajectories is obtained. Finally, the matching of the trajectory is determined by the user trajectory vector. Experimental results on BrightKite shows that the F-measure is improved by 2%～4% compared with the existing frequency based and distance based algorithm. The proposed algorithm can effectively extract the order feature of location sequence, and realize the trajectory based user identification across social networks.
Key words:Social networks/
User identification/
Trajectory similarity/
Paragraph2vec
PDF全文下载地址:
https://jeit.ac.cn/article/exportPdf?id=e8b64502-fa4d-4a9b-a479-0c575416a8c5
