Research On Technology Of Efficient Privacy-Preserving Recommendation System

The rapid development of technologies such as mobile intelligent terminals and the Internet of Things has enabled the Internet to enter an era of big data.The recommendation system can filter useful information related to users from the information overloaded network to provide personalized recommendation services.Combining the recommendation system with cloud computing technology can make the system more efficient,but the recommendation function of the recommendation system is based on the collection and analysis of users data,if the data is outsourced to a semi-trusted cloud server,then it will involve the issue of privacy leak and whether the cloud server performs the calculations correctly,so the users need to encrypt the data and verify the cloud computing results.Then,how to design a safe and efficient verifiable recommendation scheme becomes the focus of the recommendation system research.The main research contents of this paper are as follows:· This paper proposes a privacy preserving vehicle network road condition recom-mend scheme.The scheme adopts the public key encryption method to encrypt the vehicle's identity and location at the same time without encrypting the monitoring data,therefore,while ensuring the privacy of the vehicle identity and the privacy of the data location,the computational cost of encrypting the monitoring data is also e-liminated.In the scenario,the cloud server performs complex collaborative filtering calculations when it is offline,and online only needs to perform simple calculations on the current vehicle monitoring data,which helps to improve the efficiency of online road condition recommendation.· This paper proposes an efficient and verifiable privacy preserving recommendation scheme.The scheme combines symmetric full-homomorphism mapping encryption method with garbled circuits technology to realize the regression calculations on a large amount of data of user without leaking the privacy and finally achieve the recommendation purpose.In addition,the user's calculation and communication overhead in the encryption and decryption process are minimized,and the user can verify the correctness of the computation result.· This paper combines recommendation technology with cloud computing technolo-gy and encryption technology,and proposes a safe and efficient recommendation tripartite computing model which composed by the user,cloud server,and a third party.The model can meet the security requirements of user on the data,and can reduce the computational overhead of user,so that the system efficiency can be improved.