The key issues of students'T process(STP)are kernel function design and hyperparameter optimiza-tion,and the optimized hyperparameter affects directly the generalization ability of the model.In order to improve the prediction accuracy of STP,this study introduces an adaptive Bayesian optimization(BO)algorithm to opti-mize the hyperparameters of STP.Taking Benchmark questions and resonant frequency modeling of microstrip antenna as examples,the experimental comparisons are conducted with several machine learning models.This research results show that STP optimized by BO has higher fitting accuracy.
维普
万方数据