For example, Richter et al. However, this approach may have resulted in missed information on potential benefit to the model, as variability of the dataset does not necessarily correspond to differentiation/classification ability To extract prediction features, a PCA was used, as it captures the variability within the dataset.