Research On Privacy Preserving Density Peaks Clustering Method In Cloud Environment

In the development process of digitalization,virtualization and informatization,all kinds of mobile terminals and servers are producing massive data at every moment.With the increasing popularity of cloud computing,cloud computing technology provides powerful computing power in data processing.More and more enterprises store data in cloud servers to save economic costs.Data mining technology can find and extract key information from massive data.However,when data mining is carried out in the cloud computing environment,the privacy information in the data may be leaked.Therefore,how to mine useful information in data while protecting privacy has certain research value.In the research of privacy protection in clustering mining,the main work of this thesis includes the following two aspects: First,each participant needs to encrypt the privacy data,and then uploads the ciphertext to the cloud server after the security division.The cloud server analysis the ciphertext,and then returns the clustering results to the users.Second,the hybrid cloud method is used to help the users complete the secure calculation tasks,in order to solve the privacy protection in the process of clustering will bring additional computing costs to the users.Public cloud and private cloud operate together.Private cloud provides secret key,and public cloud completes clustering analysis.Relevant research results are as follows.(1)In order to solve the problem that the privacy information of users and the intermediate information produced by the cloud server in the clustering process may be leaked in the cloud computing environment,a privacy protection density peaks clustering algorithm is proposed to improve the security and availability of clustering.First,the cloud service provider calculates the cluster centers without knowing the user's specific privacy data,and does not disclose any cluster centers information to the users.Second,the cloud server allocates the users safely,and prevents each user from obtaining the privacy information of the other members in the same cluster.The cloud server can calculate the nearest clustering center for each participant without knowing the specific privacy data of the participant,and does not disclose any clustering information to the participant.A participant does not know the privacy information of the other participants in the same cluster.The security analysis and the comparative experiments show that the scheme in this thesis is safe and efficient.(2)In order to solve the problem that users need to spend a lot of computation in the process of privacy protection clustering,we propose a privacy protection density peaks clustering algorithm based on grid,which can improve the accuracy and security of clustering and reduce the user's computation cost.First,the client uses the secret key generated by private cloud and homomorphic encryption scheme to encrypt the data.Second,the client uploads the encrypted object to the public cloud,and implements a series of security protocols through the public cloud.Finally,the cloud server uses the grid idea to quickly find the clustering centers,and then returns the clustering results to the client to eliminate the perturbation.The experimental results on UCI and real datasets show that the scheme can ensure that the client has low computational complexity,and improve the efficiency and accuracy of the clustering algorithm under the premise of protecting the privacy of users.