MCDP:基于神经网络的多集群分布式差分隐私数据发布方法

大数据应用能够为人们的生活和工作方式提供便捷，但包含消费记录、社交关系、地理位置等个人隐私信息的数据在发布过程中可能被服务提供商收集，用户隐私面临巨大威胁.本文首次提出了一个基于神经网络的多集群分布式差分隐私数据发布方法，能够显著缓解单服务器的数据处理压力.同时，利用神经网络算法进行隐私参数预测明显提高了预测精度和预测效率，并且集群之间不同的隐私参数也保证了方案的灵活性.此外，由于中心服务器存储的是经过差分隐私处理后的统计数据，即使中心服务器由于遭受攻击导致存储的数据泄露，也能确保用户数据隐私.实验对比分析表明，我们的方法在隐私处理效率、隐私保护强度、预测精度和预测效率等方面都有明显优势.

MCDP:Multi-Cluster Differential Privacy Data Publishing Method Based on Neural Network

Big data applications provide convenience for people's life and work style,but in the process of data publishing,personal privacy information,such as consumption records,social relations,and so on,are collected by service providers all the time,and users' privacy is threatened greatly.Aiming at the significant relief of data processing pressure on single server,we propose a multi-cluster distributed differential privacy data publishing method based on neural network(MCDP),which effectively improves the prediction accuracy and efficiency,and different privacy parameters between clusters guarantee the flexibility of the protocol.Especially,because the central server stores statistical data after differential privacy processing,it does not collect individual privacy data,even if the central server is attacked,the user data will not be leaked.Experiments show that MCDP has obvious advantages in privacy processing efficiency,privacy protection intensity,prediction accuracy and prediction efficiency.