FIGURE 5.
(A) Top 10 principal component analysis (PCA) features and their variance ratios. (B) R-square average of 100 trainings of deep neural network (DNN) based on PCA, LASSO, and original features on the training set and test set. (C) Train set R-square distribution of DNN based on PCA, LASSO, and original features in 100 trainings. (D) The test set R-square distribution of DNN based on PCA, LASSO, and original features in 100 trainings.
