fig6
From: Prediction of arsenic (III) adsorption from aqueous solution using non-neural network algorithms
Figure 6. GPR, GPR with PCA, and GPR optimize algorithms were used with variating input features to (A) R squared; (B) RMSE; and (C) MAE. Model of GPR optimize algorithm was tested with (D) R squared; (E) RMSE; and (F) MAE. GPR: Gaussian process regression; PCA: principal component analysis; RMSE: root mean squared error; MAE: mean absolute error.
