Enterprise credit data model construction method and system based on incremental federated learning

The invention discloses an enterprise credit data model construction method and system based on incremental federal learning, and belongs to the technical field of information, and the method comprises the steps: S1, carrying out the initial training of an enterprise credit data global model; s2, on the basis of a new global model of the initial global model training and the newly added credit data, performing re-optimization training on the enterprise credit data global model; and S3, setting the obtained current corrected global model and credit data as initial parameters of next-round training, repeating the operations from the step S1 to the step S2, and performing next-round global model optimization of the enterprise credit data global model. According to the method, the incremental federation learning algorithm is used, newly-added credit data are continuously collected to serve as data samples of next model training, the precision of a local model can be effectively improved, the model is prevented fr