Fig 1. Importance of the model’s features.

Note: A: Each variable’s relative contribution is scaled with the sum of all adding to 100. B: Ratio between “permutation error” and “original error” was calculated, which means features with an importance of (around) 1 are not relevant for the model.