Fig. 1: Performances of the ML models on the glass densities of the training set.
From: Predicting densities and elastic moduli of SiO2-based glasses by machine learning
a Performance of the GBM-LASSO model. b Distribution of residuals between the GBM-LASSO predictions and the MD results of the training set. c Performance of the M5P model. d Distribution of residuals between the M5P predictions and the MD results of the training set. The curved lines in b, d are normal distributions constructed from the mean and the standard deviation of the residuals. The data points are grouped into four categories based on their glass chemistry, which are pure amorphous SiO2, type-I glasses that only contain alkane and alkane earth oxides as additives, type-II glasses that contain Al2O3 and other oxides, and type-III glasses that contain rare-earth and other oxides.
