New features include:
1: Regression models have been added to the original built-in Gaussian process model, including statistical analysis based on Bootstrap sampling regression analysis and multi-model inference. Controlled by the parameter Kriging_model.
2: Added data storage function, all forecast of training samples will be stored locally in a folder named Bgolearn.
3: Added recommended sample target value display function, supporting output of the predicted target value of recommended samples in V2.2.4.
4: Added parallel computation, multi-process computing can be performed by specifying the number of threads for the knowledge gradient function, speeding up the calculation efficiency. Controlled by the parameter Proc_num, such as Proc_num=4.