[1]华雯丽,黄 刚,唐 震.一种融合差分隐私的随机游走算法[J].计算机技术与发展,2021,31(09):112-117.[doi:10. 3969 / j. issn. 1673-629X. 2021. 09. 019]
　HUA Wen-li,HUANG Gang,TANG Zhen.A PersonalRank Walk Algorithm Fusing Differential Privacy[J].,2021,31(09):112-117.[doi:10. 3969 / j. issn. 1673-629X. 2021. 09. 019]
点击复制

一种融合差分隐私的随机游走算法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年09期
页码:
112-117
栏目:
网络与安全
出版日期:
2021-09-10

文章信息/Info

Title:
A PersonalRank Walk Algorithm Fusing Differential Privacy
文章编号:
1673-629X(2021)09-0112-06
作者:
华雯丽; 黄 刚; 唐 震
南京邮电大学 计算机学院、软件学院、网络空间安全学院,江苏 南京 210023
Author(s):
HUA Wen-li; HUANG Gang; TANG Zhen
School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China
关键词:
随机游走; 转移矩阵; 差分隐私; 拉普拉斯机制; 指数机制
Keywords:
random walk; transfer matrix; differential privacy; Laplace mechanism; exponential mechanism
分类号:
TP309
DOI:
10. 3969 / j. issn. 1673-629X. 2021. 09. 019
摘要:
如今的信息化时代,用户之间的社交网络信息越发详细,发布这些网络数据经常会威胁到一些个人隐私。 而推荐算法中,根据用户物品之间的二分图关系,进行随机游走推荐,能更可靠地推荐目标用户可能选择的物品。 由于随机游走复杂度过高,一般将图转化成转移矩阵进行计算,但是游走时无法保证该目标用户以及其他用户的隐私信息。 在隐私得不到保护的前提下,用户个人利益会受到威胁,也容易导致丢失用户的后果。 对此,需要在发布图之前处理好数据,尽量保证数据的隐私性。 而差分隐私能够在数学定义上很好地保证用户的隐私,由此在随机游走算法( PersonalRank) 的基础上,对转移矩阵通过拉普拉斯机制加噪,随机游走计算之后,再以指数机制输出推荐结果,保证了用户的信息隐私。
Abstract:
In today’s information age,social network information between users is becoming more and more detailed, and the release of such network data often threatens some personal privacy. In the recommendation algorithm, random walking recommendation is carried out according to the relationship between users and items,which can more reliably recommend the items the target user may choose. Due to the high complexity of the random walk,the graph is generally transformed into the transition matrix for calculation,but the privacy information of the target user and other users cannot be guaranteed during the walking. Without the protection of privacy,users’ personal interests will be threatened,which will easily lead to the loss of users. In this regard, it is necessary to deal with the data before publishing the graph,and try to ensure the privacy of the data. Differential privacy can guarantee users ’ privacy in mathematical definition.Therefore,on the basis of Personal Rank,the transfer matrix is denoised by Laplace mechanism. After the random walk calculation,there commendation results are output by exponential mechanism,which ensures users’ information privacy.

相似文献/References:

[1]杜阳阳,李华康,李涛.基于Node2vec 的改进算法的研究[J].计算机技术与发展,2018,28(07):6.[doi:10.3969/ j. issn.1673-629X.2018.07.002]
　DU Yang-yang,LI Hua-kang,LI Tao.Research on Improved Algorithm Based on Node2vec[J].,2018,28(09):6.[doi:10.3969/ j. issn.1673-629X.2018.07.002]
[2]李强[],陈光化[],余渊[]. 基于随机游走和混合高斯模型的运动目标检测[J].计算机技术与发展,2017,27(06):11.
　LI Qiang[],CHEN Guang-hua[],YU Yuan[]. Moving Target Detection Based on Random Walk and GaussianMixture Model[J].,2017,27(09):11.
[3]凌哲,李茂军.基于非均匀变异算子的状态空间进化算法[J].计算机技术与发展,2018,28(09):68.[doi:10.3969/ j. issn.1673-629X.2018.09.015]
　LING Zhe,LI Mao-jun.State Space Evolutionary Algorithm Based on Non-uniform Mutation Operator[J].,2018,28(09):68.[doi:10.3969/ j. issn.1673-629X.2018.09.015]
[4]李 普,陈 黎.基于超像素随机游走的自然场景图像分割方法[J].计算机技术与发展,2021,31(12):61.[doi:10. 3969 / j. issn. 1673-629X. 2021. 12. 011]
　LI Pu,CHEN Li.Natural Scene Image Segmentation Method Based on Super-pixel Random Walk[J].,2021,31(09):61.[doi:10. 3969 / j. issn. 1673-629X. 2021. 12. 011]
[5]李梓成,代永强.一种改进的鲸鱼优化算法[J].计算机技术与发展,2023,33(02):173.[doi:10. 3969 / j. issn. 1673-629X. 2023. 02. 026]
　LI Zi-cheng,DAI Yong-qiang.An Improved Whale Optimization Algorithm[J].,2023,33(09):173.[doi:10. 3969 / j. issn. 1673-629X. 2023. 02. 026]
[6]高亚星,赵旭俊,曹栩阳.基于融合数据自表示的离群点检测算法[J].计算机技术与发展,2023,33(12):41.[doi:10. 3969 / j. issn. 1673-629X. 2023. 12. 006]
　GAO Ya-xing,ZHAO Xu-jun,CAO Xu-yang.An Outlier Detection Algorithm Based on Fusion Data Self-representation[J].,2023,33(09):41.[doi:10. 3969 / j. issn. 1673-629X. 2023. 12. 006]
[7]王 烁,顾亦然,黄丽亚.基于知识子图与注意力机制的在线课程推荐模型[J].计算机技术与发展,2024,34(04):139.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 021]
　WANG Shuo,GU Yi-ran,HUANG Li-ya.An Online Course Recommendation Model Integrating Knowledge Subgraph and Attention Mechanism[J].,2024,34(09):139.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 021]

更新日期/Last Update: 2021-09-10