Figure 6. 10-fold cross-validation flow chart. In each cross-validation step, the dataset with 57 AD samples and 47 NC samples is divided into ten subsets, nine of which are used as training set and the rest as test set. In the training step, we use the training set training GSplit LBI model. In the test step, we use the trained model to predict the test set. Finally, the results of ten folds are stacked together as the results of this 10-fold cross-validation step.