INTERSECTION DATA GENERATION METHOD, AND FEDERATED MODEL TRAINING METHOD BASED ON INTERSECTION DATA

摘要

Provided is an intersection data generation method. The method comprises: receiving an ID intersection request sent by a data request terminal, wherein the ID intersection request carries at least one piece of user ID information; responding to the ID intersection request; receiving second encryption data and third encryption data; inputting first encryption data, the second encryption data and the third encryption data into a pre-configured intersection model to perform intersection processing, so as to obtain an intersection data set and a non-intersection data set; performing feature tagging processing on each piece of non-intersection data in the non-intersection data set, so as to generate a plurality of virtual features; and sending the intersection data set and the plurality of virtual features to the data request terminal to perform federated training, and then uploading the intersection data set and the plurality of virtual features to a blockchain. The method solves the problem of federated learning being prone to causing the leakage of user information, thereby improving the data security of a user.