Reduce features after calculating permutation (mda) feature importance on hierarchical clusters

Question

I have a classification task that involves feature importance using mdi, mda and shapley values. I am applying clustered mdi and mda using k-means and hierarchical clustering.