Hyper-parameter optimization method and device for credit risk control model

The invention discloses a hyper-parameter optimization method and device for a credit risk control model, relates to the field of credit risk control models and the technical field of machine learning, and mainly aims to realize that a customer selects an optimization hyper-parameter group suitable for the credit risk control model according to evaluation indexes concerned by business requirements. According to the main technical scheme, the method comprises the following steps: screening an initial hyper-parameter group set for a current sample set from hyper-parameter groups adopted by a historical task sample set on the basis of the characteristics of the optimization task and the number of hyper-parameters to be optimized, and taking the initial hyper-parameter group set as an initial hyper-parameter group set of the optimization task; performing multi-preset evaluation index optimization on hyper-parameter groups in the initial hyper-parameter group set by using a genetic algorithm to obtain an optimized