Note that for the Gini index and entropy measure, you can use an equal-width binning strategy (say, 5-10 bins) to discretize each feature. Rank the features from most important to least important according to each of the three feature importance schemes. Are there any differences?
