What's the point of compute variable importance with p-values?

[A-Pai]

What's the point of compute variable importance with p-values?--function " importance_pvalues" Is it for us to have a “natural cutoff”? which is the significance level？Just like the paper “Acomputationally fast variable importance test for random forests for high-dimensional data” said：

[mnwright]

Yes, exactly. One might argue that p-values also need a cut-off value, but at least those are easier interpretable than variable importance values.