基于本地差分隐私的网络测量Top-k目的地址估计研究  

Research on Top-k destination address estimation for network measurement based on local differential privacy in time series network

作　　者：宋绮梦 SONG Qimeng(Nanjing University of Aeronautics and Astronautics,Nanjing 210000,China)

机构地区：[1]南京航空航天大学,南京210000

出　　处：《计算机应用文摘》2024年第10期122-124,共3页Chinese Journal of Computer Application

摘　　要：针对网络测量中的Top-k目的地址问题,设计了基于布谷鸟过滤器的ε-本地差分隐私Top-k目的地址估计机制。在保护隐私的前提下,该机制能够提高对Top-k目的地址的估计效果。采用了多轮迭代方法,将布谷鸟过滤器和本地差分隐私模型结合,在服务器端和客户端进行了一轮交互。理论分析部分证明了所提出的机制符合本地差分隐私模型,并在数据集上进行了实验,对比了F1分数、NCR分数以及L1误差,验证了所提机制的有效性。In response to the Top-k destination address problem in network measurement,a cuckoo based filter was designedε-Local differential privacy Top-k destination address estimation mechanism.This mechanism can improve the estimation performance of the top k destination address while protecting privacy.We adopted a multi round iterative method,combining the cuckoo filter with a local differential privacy model,and conducted a round of interaction between the server and the client.The theoretical analysis section proves that the proposed mechanism conforms to the local differential privacy model,and experiments are conducted on the dataset to compare F1 score,NCR score,and L1 error,verifying the effectiveness of the proposed mechanism.

关 键 词：网络测量 隐私保护 Top-k目的地址 本地差分隐私 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]