An adaptive privacy protection framework for user energy data using dictionary learning and watermarking techniques

With the rise of user energy consumption data as a significant data asset, data privacy has emerged as a critical concern. To address users' diverse attitudes towards data sharing and their varied usage requirements, this paper introduces an adaptive privacy protection framework for user energy data based on dictionary learning and watermarking techniques. Central to this framework is an innovative digital watermark anonymization method designed to meet the dual objectives of encryption and anonymous data sharing. This method employs sparse dictionary decomposition to embed confidential user information within sparse coefficients, significantly enhancing computational efficiency while minimally impacting the integrity of the original data. Additionally, through sparse data representation, the framework achieves effective data compression, addressing the challenges of extensive storage requirements inherent in maintaining original energy consumption data, encryption results, and the data sharing process. Security analysis and case studies demonstrate the proposed method's robustness against eavesdropping, unauthorized access, and other security threats.