Federal learning security aggregation method and system for distributed load prediction

The invention provides a federated learning security aggregation method and system for distributed load prediction, and the method comprises the steps: employing an LSTM algorithm model as a basic training model, enabling a participant to receive an initialized model parameter from a central server, and training a local model through employing local data; the participants upload the trained and updated local model parameters to a central server; and the central server determines approximate global model parameters based on model parameter similarity, aggregates the global model parameters by using a distance-based weighting method, and then allocates the global model parameters to the participant local models for next round of training. According to the method and the system provided by the invention, each load participant does not need to share local data, so that the privacy of sensitive information is protected; the interference of the malicious local model on the global model is eliminated, and the sample