| 引用本文： |
-
薛晖,孙波,司成祥,张伟,房婧.跨社交网络用户身份链接回顾与展望[J].信息安全学报,2026,11(2):300-312 [点击复制]
- XUE Hui,SUN Bo,SI Chengxiang,ZHANG Wei,FANG Jing.Advance in user identity linkage across online social networks[J].Journal of Cyber Security,2026,11(2):300-312 [点击复制]
|
|
| |
|
|
| 本文已被：浏览 10次 下载 4次 |
码上扫一扫！ |
| 跨社交网络用户身份链接回顾与展望 |
|
薛晖1,2, 孙波1,3, 司成祥3, 张伟3, 房婧3
|
|
|
| (1.中国科学院信息工程研究所 北京 中国 100093;2.中国科学院大学网络空间安全学院 北京 中国 100049;3.国家互联网应急中心 北京 中国 100029) |
|
| 摘要: |
| 随着互联网的飞速发展,社交网络平台(又称在线社交网络)也日益普及和多样化,为了更好地利用每个社交网络平台提供的服务,用户往往会加入多个社交网络平台。链接同一个自然人在不同社交网络平台中的账户,称为用户身份链接。通过用户身份链接可以充分了解用户的兴趣,极大地丰富用户画像,进而用于数字营销和推荐系统。本文首先通过回顾用户身份链接方法在发展过程中所使用的不同特征类型,提出了一种用户身份链接问题的通用形式化定义,适用于属性、网络、内容、行为等各种特征类型及其任意组合。然后根据用户身份链接的特征提取和模型构建两个阶段对现有用户身份链接方法进行了分类分析,并分别从性能、计算开销、鲁棒性维度对各类方法进行了比较和评价。而后分析了现有方法使用的不同数据集和评价指标,说明了数据集的主要获取方式,并给出了目前用户身份链接领域无公开公认的基准数据集的原因。最后讨论了用户身份链接存在的问题与挑战,展望了用户身份链接的未来研究趋势。本文通过提出一种用户身份链接问题的通用定义、比较分析已有用户身份链接方法、讨论存在的问题和展望未来研究趋势,将用户身份链接问题的现状和未来以清晰的结构化的方式进行分析展示,有助于研究人员对该领域的相关研究形成系统性的理解和把握,进而做出更加深入的研究工作。 |
| 关键词: 社交网络 用户身份链接 账号关联 锚链接预测 用户画像 |
| DOI：10.19363/J.cnki.cn10-1380/tn.2026.03.19 |
| 投稿时间：2020-12-30修订日期：2021-03-15 |
| 基金项目: |
|
| Advance in user identity linkage across online social networks |
|
XUE Hui1,2, SUN Bo1,3, SI Chengxiang3, ZHANG Wei3, FANG Jing3
|
| (1.Institute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, China;2.School of Cyber Security, University of Chinese Academy of Sciences, Beijing 100049, China;3.National Internet Emergency Center, Beijing 100029, China) |
| Abstract: |
| With the rapid development of the Internet, social network platforms (also known as online social networks) have become increasingly popular and diversified. In order to make better use of the services provided by each social network platform, users often join multiple social network platforms. Linking the accounts of the same natural person in different social network platforms is called user identity linkage. Through user identity linkage, we can fully understand the user's interests, and greatly enrich the user portrait, which is used in digital marketing and recommendation system. In this paper, by reviewing the different feature types used in the development of the user identity linkage method, a general formal definition of the user identity linkage problem is proposed, which can be applied to various feature types such as attribute, network, content, behavior and any combination of them. Then, according to the two stages of feature extraction and model construction of user identity linkage, the existing user identity linkage methods are classified and analyzed, and different methods are compared and evaluated in terms of performance, computing cost and robustness. Then, different datasets and evaluation indicators used by existing methods are analyzed, the main methods of obtaining datasets are explained, and the reason why there is no publicly recognized benchmark datasets in the field of user identity linkage is given. Finally, the problems and challenges of user identity linkage are discussed, and the future research trend of user identity linkage is forecasted. By proposing a general definition of user identity linkage problem, comparing and analyzing existing user identity linkage methods, discussing existing problems and looking forward to future research trends, this paper analyzes and presents the current situation and future of user identity linkage problem in a clear and structured way, which helps researchers to form a systematic understanding and grasp of related research in this field, and then make more in-depth research work. |
| Key words: online social network user identity linkage account association anchor link prediction user profile |
|
|
|
|
|
