From: Intervention in prediction measure: a new approach to assessing variable importance for random forests
Variables
Y 1|X 1=0
Y 1|X 1=1
Y 2|X 1=0
Y 2|X 1=1
Scenario 3
2 + 2 ·X 2 + N(0,0.1)
3 + 4 ·X 2 + N(0,0.2)
2 + 3 ·X 2 + N(0,0.15)
3 + 5 ·X 2 + N(0,0.2)
Scenario 4
idem
4 + N(0,0.2)
5 + N(0,0.2)