Big data security risk control model based on federated learning algorithm

Abstract

The distributed big data security risk control model achieves the control of big data security risk by distributed training of data feature vectors. The lack of processing of encrypted data leads to weak generalization ability. In this regard, a big data security risk control model based on federal learning algorithm is proposed. The heterogeneous data is formatted and the original data is preprocessed by data discretization and data scaling. The optimized federation learning algorithm is used to match the encrypted data, and the big data security risk control model is constructed to improve the generalization ability of the model. In the experiments, the proposed model is tested for its generalization ability. The analysis of the experimental results shows that the big data security risk control model constructed by using the proposed method has high data generalization ability.