“…The offline SBC-RBFNN model structure is obtained by combining ( 9)- (11), denoted by y = h(x). To obtain the parameters for one individual model, including the weights W [1] , W [2] , and W [3] , the biases b [1] , b [2] , and b [3] , as well as the center vector [z c1 , z c2 , • • • , z cn ], the datasets are randomly split into the base model group, training group, and testing group, and they are used for base model generation, SBC-RBFNN model training, and model verification, respectively. The RMSE between the model and the expected output, calculated by (8), is used as the loss function in the training process.…”