This repository is a working repo for the Hyperparameter optimization of surrogate modeling for structural design (HOpt4SMSD) using machine learning algorithms: Gaussian process regression (GPR), support vector machine (SVM), random forest regression (RFR), and artificial intelligent network (ANN).
(1)Pre-requisits:
- The R language version 3 is required.
- Several R packages are required ro run the codes: mlr and its required packages, mxnet, mlrMBO, readr, smoof;
(2)The details can refer to:
Du, Xianping, Hongyi Xu, and Feng Zhu. "Understanding the effect of hyperparameter optimization on machine learning models for structure design problems." Computer-Aided Design 135 (2021): 103013.