Compute embedded estimates of input feature importance
imp = varimportance(t)
imp = varimportance(t) computes estimates of input feature importance for tree t by summing changes in the risk due to splits on every feature. The returned vector imp has one element for each input variable in the data used to train this tree. At each node, the risk is estimated as node impurity if impurity was used to split nodes and node error otherwise. This risk is weighted by the node probability. Variable importance associated with this split is computed as the difference between the risk for the parent node and the total risk for the two children.