Face to Electricity Data Transmission of Composite Differential Private Recommend Method Research

Abstract

With the construction and development of China’s smart power grid, it has realized the power information interconnection, but also realized the collection of electricity user information fine, to bring convenience to users, but also led to the risk of privacy leakage. The traditional method of privacy protection has certain limitations to the protection of users’ privacy data: only to a certain extent to protect the privacy of users. With the advent of advanced technologies such as machine learning, the ability of attackers to speculate on privacy has improved significantly, and traditional methods of privacy protection have been difficult to work with. This paper summarizes and analyzes the centralized differential privacy method and the localized differential privacy method in smart grid data transmission. The characteristics and advantages of The Laplace mechanism, Gaussian mechanism, and index mechanism are analyzed and compared on the addition of noise disturbance mechanism in differential privacy method. In addition, this paper introduces the current researchers on the local differential privacy methods and noise-making mechanism improvement methods. Finally, a K-Means user clustering method based on K-Means is proposed, the main method is to do K-Means clustering analysis based on the sensitivity of different user groups, and then use different differential privacy methods according to different group categories.