A Data Mining Technology of Electricity Trading Data Based on Data Sandbox

With the rapid rise of data mining technology such as machine learning, the flow of data production factors has become a key issue. Due to the high added value and high sensitivity, the mining and utilization of electricity trading data needs security protection. This paper analyzes the value increment of electricity trading data, the security risks and limitations in current applications. Furthermore, this paper proposes a new data mining scheme for electricity trading. Technically, this paper divides the mining tasks for electricity trading data into three categories, and designs a subsystem module for each category of task. Compared with the traditional scheme, this scheme can ensure that users can complete data mining tasks without touching the real data, such that it can support the opening and utilization of any electricity trading data. Compared with the traditional scheme, this scheme greatly increases the scope of data opening. Compared with the privacy protection technology such as secure multi-party computing, our scheme only requires plaintext operations on electricity trading data, and does not require complex ciphertext computation, and hence the computational and communication costs are significantly reduced.