Fig. 4

Flowchart of the Machine Learning process used to assess the performance of each algorithm tested. Each dataset is initially split into two subsets: the ‘Learning dataset’ and the ‘Independent test set’. Subsequently, the former undergoes a 5-fold cross validation strategy, where Training sets are used to select informative features (‘Feature Selection’) and Validation sets to test the classifier performance (‘Evaluation’). Finally, the Best Model is selected and, then, assessed on the Independent test set (‘Evaluation’): the last evaluation step is used to compare the performance of each feature selection method