Table 2 Regression results of the GBM-LASSO and M5P machine learning models on the training set, including root mean squared error (RMSE), and the percentage of predictions within 5%, 10%, 20%, and 30% relative errors according to Eq. 6, respectively.
From: Predicting densities and elastic moduli of SiO2-based glasses by machine learning
Property | Model | RMSE | Percent of predictions within relative error of | |||
|---|---|---|---|---|---|---|
2.5% | 5% | 10% | 20% | |||
Density | GBM-LASSO | 0.0229 | 98.8 | 100.0 | 100.0 | 100.0 |
M5P | 0.0325 | 96.6 | 100.0 | 100.0 | 100.0 | |
K | GBM-LASSO | 2.99 | 33.9 | 61.8 | 91.0 | 99.6 |
M5P | 2.59 | 40.6 | 70.1 | 94.6 | 99.6 | |
G | GBM-LASSO | 1.31 | 47.4 | 76.3 | 96.0 | 100.0 |
M5P | 0.97 | 57.6 | 89.8 | 99.4 | 100.0 | |
