Construction of a multi-technology fusion e-commerce data transaction optimization model based on federated learning

The constraints of data protection make the data confined to different enterprises and organizations, forming many “data islands” and making it difficult to bring out the important value it contains. In this paper, we use federated learning technology as the service foundation and introduce differential privacy, federation chain, interstellar file system, and trusted execution environment to construct a multi-technology fusion e-commerce data transaction method. The three concepts of budget feasibility, individual rationality, and incentive mechanisms are applied to the data transaction scenario to design smart contracts. At the same time, the incentive mechanism is created by combining the trusted execution environment and Shapley value, and the transaction process model is optimized. Simulation comparison is carried out based on the public dataset of the Taobao e-commerce platform, and the experimental results show that MTFDT can realize the accurate evaluation of the model training effect, and the incremental profit stabilization point of the data buyer and seller is around 0.4, which improves the fairness of benefit distribution.