关键词: 数据可信共享, 流模型, 主成分分析, 同态加密, 隐私保护, 智能电网

Abstract: under the context of deepening reforms in the electricity market, market entities generate massive amounts of high-dimensional sensitive data during their participation in electricity trading and operations. These data hold significant market value, but their sharing and collaborative analysis face severe challenges such as the leakage of core commercial secrets and the infringement of user privacy, which severely hinder improvements in market efficiency and the efficient participation of distributed resources. Existing privacy protection schemes (such as multi-party secure computation and differential privacy) suffer from drawbacks such as high computational overhead or significant loss of data utility, making them unable to meet the dual requirements of real-time processing and data fidelity demanded by the high-frequency transactions and detailed analysis in the power market. To address this issue, a data trusted sharing scheme based on stream models is proposed, which integrates the reversible characteristics of stream models, the powerful dimensionality reduction capabilities of principal component analysis (PCA), and the privacy protection advantages of homomorphic encryption. Specifically, the stream model is first used to map the original data to a latent space for preliminary transformation, followed by PCA technology for effective dimensionality reduction, and finally homomorphic encryption to comprehensively protect data privacy. Experimental results show that the proposed scheme effectively reduces pixel correlation—by an average of 48.09 % compared with the original image—thereby enhancing both the security of the encrypted data and the efficiency of the encryption computation, cutting processing time by approximately 89.6 %.

Key words: secure data sharing, flow model, principal component analysis, homomorphic encryption, privacy protection, smart grid