Table 2. GPR Training Result with Different Kernel Functionsa.
RMSE | R2 | MSE | MAE | |
---|---|---|---|---|
GPR with the squared exponential kernel | 16.287 | 0.97 | 265.27 | 5.8231 |
GPR with the rational quadratic kernel | 7.9321 | 0.99 | 62.919 | 3.2494 |
GPR with the combined kernel | 1.8060 | 1.00 | 3.262 | 1.2807 |
GPU is used for parallel computing.